makemandb.c: spell "metadata" consistently

makemandb.c: fix grammar in a comment

fix various typos in comments and makefs README file.

makemandb/*: fix spelling of database and consistency of SQLite

Don't index outside the mdocs array of function pointers. Analysis and
suggested fixes from Tom Lane. I played it safe and went with (my
variation of) the minimal fix.

port-hppa/56118: sporadic app crashes in HPPA -current

makemandb.c: fail sooner if man page dirs can't be found

There's no point initializing database state if we're then going to
fail to locate any man page sources. Make all the initial state checks
contiguous for simplicity and readability. Also, free the variable
"command" on the error path, and correct the error message.

s/trival/trivial/ in comment.

branches: 1.60.2;
PR misc/54213: Fix performance of whatis(1) when no matches are found

In revision 1.6 of whatis.c the query was modified to return matches for names found
in MLINKS of the man pages as well. However it was slow. The reason probably being that it
required a join. But more importantly the where condition on an FTS virtual table column
is very slow. To avoid the join and the expensive where condition on the virtual table,
add the name_desc column to the mandb_links table as well. This improves the performance
of whatis(1) to the original level at the expense of slight data duplication.

Bump the schema to force database rebuild to take account for the new column addition

remove unneeded header.

adjust to the new mandoc api

Adjust makemandb for the latest mandoc

ok christos@

In the latest mandoc (mdocml) the signature (prototype) of
mparse_alloc() altered - update the call here to compensate.

This fixes the build (of makemandb), but I am not sure that
the changed version is what is desired - someone who knows
something about all of this should validate ... I just copied
the invocation from mandoc's demandoc.c (which seems likely
to be at least a similar kind of usage).

branches: 1.55.8; 1.55.10;
Get rid of unnecessary variable.

We do need to copy the return value from dirname(3) since there it is a static
buffer and can be overwritten in between. I overzealously removed this in one
of my previous commits.

Avoid dereferencing pointer at multiple places, instead use a local variable.

Remove the table name parameter from the check_md5 function.

There is only one table storing the md5 checksums, so we can hardcode the table
name instead of passing it as a function argument.

Avoid copying strings where it is not needed.

Avoid a call to strncmp when comparing only the first character of the string.

Bring the comment in sync with code (after changes brought by the last commit).

Don't parse Nm macro when it occurs anywhere outside the NAME section.

mandoc(3) already generates the text node representing the value for the .Nm macro.
Doing our own parsing for .Nm on top of that leads to large duplication of text
in the database. This gets specially worse for man pages with large NAME sections,
such as queue(3).

branches: 1.47.2;
Use libarchive 3.x interface and not obsolete 2.x versions.

branches: 1.46.2;
Escape hyphen when parsing .Nd

Don't ignore symlinks.
There can be symlinks which are pointing to man pages not installed in
one of the _default locations mentioned in man.conf or MANPATH. For example
there are man pages in /usr/pkg/man which are symlinked to pages in
/usr/pkg/lib/perl5/man. If we ignore symlinks, we would not be able to
index such pages installed outside the default set of directories.

(Also, the symlink test was incorecct, so we never noticed this issue)

Ok christos@, wiz@

We don't need to parse the sections we don't index, so stop early. Saves few
instructions.

With the latest release of mandoc, makemandb(8) started to parse some
sections multiple times. This started to happen because, pmdoc_Sh(), the handler function
responsible for parsing the Sh macros, used to recursively go through all the child
nodes and then the next nodes starting from top level Sh block node.
Now, once it has processed all the child nodes of the top level block node,
it moves to the next node, which is the top level block node of the next section and
in this way one call to pmdoc_Sh() was causing a complete pass through the
man page. Since, mandoc(3) calls pmdoc_Sh() for each .Sh macro in the man
page, it would result in parsing some of the sections multiple times.
This never happened with the previous versions of mandoc, so we never noticed.

I've fixed this by starting the parse sequence of the Sh macro from its body, which gurantees
that we will stop once that section ends.

ok christos@

Use deroff() from mandoc(3) to directly parse the Nd macro rather
than parsing it by hand.

With the latest mandoc(3), the .Nd macro was getting parsed twice. This fixes
that problem and cleans up the code as well.

ok christos@

Fix extraction of section number and machine architecture from man pages.
With the latest API, mdoc_validate()/man_validate() needs to be called before
reading the roff_man.meta field, otherwise it is NULL.

Also, if a man page doesn't specify machine architecture, don't default to '?'
, let it be stored as null in the db. Otherwise, the output of apropos(1) shows
the names of the results as \?/<title>

Sync with API changes.

branches: 1.39.2;
Avoid possible buffer overflow while parsing NAME section of man(7) pages.
Also, simplify copyging of strings, use estrdupn instead of emalloc + memcpy.

Patch from christos@, XXX comment by me

Reuse variable from previous line.

PR/51062: Abhinav Upadhyay: Allow non numeric sections to be indexed and
searched by apropos(1).
Fold long lines.

PR/51040: Abhinav Upadhyay: Fix memory leak

PR/51039: Abhinav Upadhyay: Check for return value of chdir(2)

PR/51034: Abhinav Upadhyay: Close database connection when failed to commit

PR/51034: Abhinav Upadhyay: makemandb(8): Close database connection when
failed to commit

PR/51006: Abhinav Upadhyay: makemandb(8) should parse escape sequences
in the NAME section

Don't crash if we have a missing section.

Adjust to the new mdocml

largely apply patch from PR bin/47392 by Abhinav Upadhyay

change some comments to reflect reality, a variable name to enhance
readability, and adds an assert for safety.

MDOC_MAX is a valid token if the type is text. Adjust.

- handle section numbers that are not single digits
- don't allocate and free needlessly

Explicitly deal with end of lists. PR 49708.

src is too big these days to tolerate superfluous apostrophes. It's
"its", people!

branches: 1.24.2;
Fix a bug that caused an error about a UNIQUE constraint violation.
Patch from Abhinav Upadhyay.

Replace non-breaking space with hyphen, and call hyphen replacement
from one more place.
Improves 'man -k midi' output.

From Abhinav Upadhyay.

branches: 1.22.2;
in update_db(), extract the full list of files to update from the db
before actually updating anything, since changing the db while the query
that extracts the list of files is still in progress results in
the extraction query failing before it finds everything.

Sync with interface change in mdocml 1.12.3.

Skip files of size 0 from indexing.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

avoid stdio assertion, failing later

remove trailing whitespace

Since mdocml decided to name headers that conflict with system ones (term.h)
move the header inclusion one up.

If you cannot parse .SH NAME, like in the case of the ksh93 man page
where the .SH is followed by a conditional:

.SH NAME
.if \nZ=0 \{\
text text text
.\}

at least don't core-dump.

Make mandb path configurable. makemandb (and related tools) use
the path from the _mandb variable from man.conf now.

Set _mandb in man.conf to same value as was used before.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

branches: 1.14.2;
Use emalloc in one more place, like the rest of the code does.
From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Add -Q flag:
Print only fatal error messages (i.e., when the database is left in
an inconsistent state and needs manual intervention).

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Sync usage with manpage.

Bug fix for PR 46733:
> makemandb always reports the same number for "Total Number of new or
> updated pages enountered" and "Total number of (hard or symbolic)
> links found".

Patch from Abhinav Upadhyay.

Fix typo in a message.

PR 46419 by Abhinav Upadhyay using his updated patch:
Clean up after removing man page aliases.

The new apropos(1) incorrectly displays hyphens in the first line
of the search results for a few man pages (for man(7) based man
pages).

Use patch from Abhinav Upadhyay in PR 46408 to fix this.

branches: 1.7.2;
Fix inverted condition when handling stale entries.
From Abhinav Upadhyay.

Expand workaround for .so usage to do the chdir call just before
starting parsing, not during the tree iteration. This gives it a chance
to work.

Add support for compressed man pages in all the usual formats.

Also handle hyphen replacement if it was used as plain input and no
backslash sequence was used at all in the line.

Be a bit more friendly to man pages using the roff .so command by
changing the current directory to the parent of the man -p entry, e.g.
/usr/share/man for /usr/share/man1.

branches: 1.2.2;
Fix C&P error with $NetBSD$

Import the new apropos/whatis.

This code has been developed by Abhinav Upadhyay as part of Google's Summer
of Code 2011. It uses libmandoc to parse man pages and builds a Full
Text Index in a SQLite database. The combination of indexing the full
manual page, filtering out stop words and ranking individual matches
based on the section gives a much improved user experience.

The old makewhatis and friends are kept under MKMAKEMANDB=no for now.

makemandb.c: fix grammar in a comment

fix various typos in comments and makefs README file.

makemandb/*: fix spelling of database and consistency of SQLite

Don't index outside the mdocs array of function pointers. Analysis and
suggested fixes from Tom Lane. I played it safe and went with (my
variation of) the minimal fix.

port-hppa/56118: sporadic app crashes in HPPA -current

makemandb.c: fail sooner if man page dirs can't be found

There's no point initializing database state if we're then going to
fail to locate any man page sources. Make all the initial state checks
contiguous for simplicity and readability. Also, free the variable
"command" on the error path, and correct the error message.

s/trival/trivial/ in comment.

branches: 1.60.2;
PR misc/54213: Fix performance of whatis(1) when no matches are found

In revision 1.6 of whatis.c the query was modified to return matches for names found
in MLINKS of the man pages as well. However it was slow. The reason probably being that it
required a join. But more importantly the where condition on an FTS virtual table column
is very slow. To avoid the join and the expensive where condition on the virtual table,
add the name_desc column to the mandb_links table as well. This improves the performance
of whatis(1) to the original level at the expense of slight data duplication.

Bump the schema to force database rebuild to take account for the new column addition

remove unneeded header.

adjust to the new mandoc api

Adjust makemandb for the latest mandoc

ok christos@

In the latest mandoc (mdocml) the signature (prototype) of
mparse_alloc() altered - update the call here to compensate.

This fixes the build (of makemandb), but I am not sure that
the changed version is what is desired - someone who knows
something about all of this should validate ... I just copied
the invocation from mandoc's demandoc.c (which seems likely
to be at least a similar kind of usage).

branches: 1.55.8; 1.55.10;
Get rid of unnecessary variable.

We do need to copy the return value from dirname(3) since there it is a static
buffer and can be overwritten in between. I overzealously removed this in one
of my previous commits.

Avoid dereferencing pointer at multiple places, instead use a local variable.

Remove the table name parameter from the check_md5 function.

There is only one table storing the md5 checksums, so we can hardcode the table
name instead of passing it as a function argument.

Avoid copying strings where it is not needed.

Avoid a call to strncmp when comparing only the first character of the string.

Bring the comment in sync with code (after changes brought by the last commit).

Don't parse Nm macro when it occurs anywhere outside the NAME section.

mandoc(3) already generates the text node representing the value for the .Nm macro.
Doing our own parsing for .Nm on top of that leads to large duplication of text
in the database. This gets specially worse for man pages with large NAME sections,
such as queue(3).

branches: 1.47.2;
Use libarchive 3.x interface and not obsolete 2.x versions.

branches: 1.46.2;
Escape hyphen when parsing .Nd

Don't ignore symlinks.
There can be symlinks which are pointing to man pages not installed in
one of the _default locations mentioned in man.conf or MANPATH. For example
there are man pages in /usr/pkg/man which are symlinked to pages in
/usr/pkg/lib/perl5/man. If we ignore symlinks, we would not be able to
index such pages installed outside the default set of directories.

(Also, the symlink test was incorecct, so we never noticed this issue)

Ok christos@, wiz@

We don't need to parse the sections we don't index, so stop early. Saves few
instructions.

With the latest release of mandoc, makemandb(8) started to parse some
sections multiple times. This started to happen because, pmdoc_Sh(), the handler function
responsible for parsing the Sh macros, used to recursively go through all the child
nodes and then the next nodes starting from top level Sh block node.
Now, once it has processed all the child nodes of the top level block node,
it moves to the next node, which is the top level block node of the next section and
in this way one call to pmdoc_Sh() was causing a complete pass through the
man page. Since, mandoc(3) calls pmdoc_Sh() for each .Sh macro in the man
page, it would result in parsing some of the sections multiple times.
This never happened with the previous versions of mandoc, so we never noticed.

I've fixed this by starting the parse sequence of the Sh macro from its body, which gurantees
that we will stop once that section ends.

ok christos@

Use deroff() from mandoc(3) to directly parse the Nd macro rather
than parsing it by hand.

With the latest mandoc(3), the .Nd macro was getting parsed twice. This fixes
that problem and cleans up the code as well.

ok christos@

Fix extraction of section number and machine architecture from man pages.
With the latest API, mdoc_validate()/man_validate() needs to be called before
reading the roff_man.meta field, otherwise it is NULL.

Also, if a man page doesn't specify machine architecture, don't default to '?'
, let it be stored as null in the db. Otherwise, the output of apropos(1) shows
the names of the results as \?/<title>

Sync with API changes.

branches: 1.39.2;
Avoid possible buffer overflow while parsing NAME section of man(7) pages.
Also, simplify copyging of strings, use estrdupn instead of emalloc + memcpy.

Patch from christos@, XXX comment by me

Reuse variable from previous line.

PR/51062: Abhinav Upadhyay: Allow non numeric sections to be indexed and
searched by apropos(1).
Fold long lines.

PR/51040: Abhinav Upadhyay: Fix memory leak

PR/51039: Abhinav Upadhyay: Check for return value of chdir(2)

PR/51034: Abhinav Upadhyay: Close database connection when failed to commit

PR/51034: Abhinav Upadhyay: makemandb(8): Close database connection when
failed to commit

PR/51006: Abhinav Upadhyay: makemandb(8) should parse escape sequences
in the NAME section

Don't crash if we have a missing section.

Adjust to the new mdocml

largely apply patch from PR bin/47392 by Abhinav Upadhyay

change some comments to reflect reality, a variable name to enhance
readability, and adds an assert for safety.

MDOC_MAX is a valid token if the type is text. Adjust.

- handle section numbers that are not single digits
- don't allocate and free needlessly

Explicitly deal with end of lists. PR 49708.

src is too big these days to tolerate superfluous apostrophes. It's
"its", people!

branches: 1.24.2;
Fix a bug that caused an error about a UNIQUE constraint violation.
Patch from Abhinav Upadhyay.

Replace non-breaking space with hyphen, and call hyphen replacement
from one more place.
Improves 'man -k midi' output.

From Abhinav Upadhyay.

branches: 1.22.2;
in update_db(), extract the full list of files to update from the db
before actually updating anything, since changing the db while the query
that extracts the list of files is still in progress results in
the extraction query failing before it finds everything.

Sync with interface change in mdocml 1.12.3.

Skip files of size 0 from indexing.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

avoid stdio assertion, failing later

remove trailing whitespace

Since mdocml decided to name headers that conflict with system ones (term.h)
move the header inclusion one up.

If you cannot parse .SH NAME, like in the case of the ksh93 man page
where the .SH is followed by a conditional:

.SH NAME
.if \nZ=0 \{\
text text text
.\}

at least don't core-dump.

Make mandb path configurable. makemandb (and related tools) use
the path from the _mandb variable from man.conf now.

Set _mandb in man.conf to same value as was used before.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

branches: 1.14.2;
Use emalloc in one more place, like the rest of the code does.
From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Add -Q flag:
Print only fatal error messages (i.e., when the database is left in
an inconsistent state and needs manual intervention).

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Sync usage with manpage.

Bug fix for PR 46733:
> makemandb always reports the same number for "Total Number of new or
> updated pages enountered" and "Total number of (hard or symbolic)
> links found".

Patch from Abhinav Upadhyay.

Fix typo in a message.

PR 46419 by Abhinav Upadhyay using his updated patch:
Clean up after removing man page aliases.

The new apropos(1) incorrectly displays hyphens in the first line
of the search results for a few man pages (for man(7) based man
pages).

Use patch from Abhinav Upadhyay in PR 46408 to fix this.

branches: 1.7.2;
Fix inverted condition when handling stale entries.
From Abhinav Upadhyay.

Expand workaround for .so usage to do the chdir call just before
starting parsing, not during the tree iteration. This gives it a chance
to work.

Add support for compressed man pages in all the usual formats.

Also handle hyphen replacement if it was used as plain input and no
backslash sequence was used at all in the line.

Be a bit more friendly to man pages using the roff .so command by
changing the current directory to the parent of the man -p entry, e.g.
/usr/share/man for /usr/share/man1.

branches: 1.2.2;
Fix C&P error with $NetBSD$

Import the new apropos/whatis.

This code has been developed by Abhinav Upadhyay as part of Google's Summer
of Code 2011. It uses libmandoc to parse man pages and builds a Full
Text Index in a SQLite database. The combination of indexing the full
manual page, filtering out stop words and ranking individual matches
based on the section gives a much improved user experience.

The old makewhatis and friends are kept under MKMAKEMANDB=no for now.

fix various typos in comments and makefs README file.

makemandb/*: fix spelling of database and consistency of SQLite

Don't index outside the mdocs array of function pointers. Analysis and
suggested fixes from Tom Lane. I played it safe and went with (my
variation of) the minimal fix.

port-hppa/56118: sporadic app crashes in HPPA -current

makemandb.c: fail sooner if man page dirs can't be found

There's no point initializing database state if we're then going to
fail to locate any man page sources. Make all the initial state checks
contiguous for simplicity and readability. Also, free the variable
"command" on the error path, and correct the error message.

s/trival/trivial/ in comment.

branches: 1.60.2;
PR misc/54213: Fix performance of whatis(1) when no matches are found

In revision 1.6 of whatis.c the query was modified to return matches for names found
in MLINKS of the man pages as well. However it was slow. The reason probably being that it
required a join. But more importantly the where condition on an FTS virtual table column
is very slow. To avoid the join and the expensive where condition on the virtual table,
add the name_desc column to the mandb_links table as well. This improves the performance
of whatis(1) to the original level at the expense of slight data duplication.

Bump the schema to force database rebuild to take account for the new column addition

remove unneeded header.

adjust to the new mandoc api

Adjust makemandb for the latest mandoc

ok christos@

In the latest mandoc (mdocml) the signature (prototype) of
mparse_alloc() altered - update the call here to compensate.

This fixes the build (of makemandb), but I am not sure that
the changed version is what is desired - someone who knows
something about all of this should validate ... I just copied
the invocation from mandoc's demandoc.c (which seems likely
to be at least a similar kind of usage).

branches: 1.55.8; 1.55.10;
Get rid of unnecessary variable.

We do need to copy the return value from dirname(3) since there it is a static
buffer and can be overwritten in between. I overzealously removed this in one
of my previous commits.

Avoid dereferencing pointer at multiple places, instead use a local variable.

Remove the table name parameter from the check_md5 function.

There is only one table storing the md5 checksums, so we can hardcode the table
name instead of passing it as a function argument.

Avoid copying strings where it is not needed.

Avoid a call to strncmp when comparing only the first character of the string.

Bring the comment in sync with code (after changes brought by the last commit).

Don't parse Nm macro when it occurs anywhere outside the NAME section.

mandoc(3) already generates the text node representing the value for the .Nm macro.
Doing our own parsing for .Nm on top of that leads to large duplication of text
in the database. This gets specially worse for man pages with large NAME sections,
such as queue(3).

branches: 1.47.2;
Use libarchive 3.x interface and not obsolete 2.x versions.

branches: 1.46.2;
Escape hyphen when parsing .Nd

Don't ignore symlinks.
There can be symlinks which are pointing to man pages not installed in
one of the _default locations mentioned in man.conf or MANPATH. For example
there are man pages in /usr/pkg/man which are symlinked to pages in
/usr/pkg/lib/perl5/man. If we ignore symlinks, we would not be able to
index such pages installed outside the default set of directories.

(Also, the symlink test was incorecct, so we never noticed this issue)

Ok christos@, wiz@

We don't need to parse the sections we don't index, so stop early. Saves few
instructions.

With the latest release of mandoc, makemandb(8) started to parse some
sections multiple times. This started to happen because, pmdoc_Sh(), the handler function
responsible for parsing the Sh macros, used to recursively go through all the child
nodes and then the next nodes starting from top level Sh block node.
Now, once it has processed all the child nodes of the top level block node,
it moves to the next node, which is the top level block node of the next section and
in this way one call to pmdoc_Sh() was causing a complete pass through the
man page. Since, mandoc(3) calls pmdoc_Sh() for each .Sh macro in the man
page, it would result in parsing some of the sections multiple times.
This never happened with the previous versions of mandoc, so we never noticed.

I've fixed this by starting the parse sequence of the Sh macro from its body, which gurantees
that we will stop once that section ends.

ok christos@

Use deroff() from mandoc(3) to directly parse the Nd macro rather
than parsing it by hand.

With the latest mandoc(3), the .Nd macro was getting parsed twice. This fixes
that problem and cleans up the code as well.

ok christos@

Fix extraction of section number and machine architecture from man pages.
With the latest API, mdoc_validate()/man_validate() needs to be called before
reading the roff_man.meta field, otherwise it is NULL.

Also, if a man page doesn't specify machine architecture, don't default to '?'
, let it be stored as null in the db. Otherwise, the output of apropos(1) shows
the names of the results as \?/<title>

Sync with API changes.

branches: 1.39.2;
Avoid possible buffer overflow while parsing NAME section of man(7) pages.
Also, simplify copyging of strings, use estrdupn instead of emalloc + memcpy.

Patch from christos@, XXX comment by me

Reuse variable from previous line.

PR/51062: Abhinav Upadhyay: Allow non numeric sections to be indexed and
searched by apropos(1).
Fold long lines.

PR/51040: Abhinav Upadhyay: Fix memory leak

PR/51039: Abhinav Upadhyay: Check for return value of chdir(2)

PR/51034: Abhinav Upadhyay: Close database connection when failed to commit

PR/51034: Abhinav Upadhyay: makemandb(8): Close database connection when
failed to commit

PR/51006: Abhinav Upadhyay: makemandb(8) should parse escape sequences
in the NAME section

Don't crash if we have a missing section.

Adjust to the new mdocml

largely apply patch from PR bin/47392 by Abhinav Upadhyay

change some comments to reflect reality, a variable name to enhance
readability, and adds an assert for safety.

MDOC_MAX is a valid token if the type is text. Adjust.

- handle section numbers that are not single digits
- don't allocate and free needlessly

Explicitly deal with end of lists. PR 49708.

src is too big these days to tolerate superfluous apostrophes. It's
"its", people!

branches: 1.24.2;
Fix a bug that caused an error about a UNIQUE constraint violation.
Patch from Abhinav Upadhyay.

Replace non-breaking space with hyphen, and call hyphen replacement
from one more place.
Improves 'man -k midi' output.

From Abhinav Upadhyay.

branches: 1.22.2;
in update_db(), extract the full list of files to update from the db
before actually updating anything, since changing the db while the query
that extracts the list of files is still in progress results in
the extraction query failing before it finds everything.

Sync with interface change in mdocml 1.12.3.

Skip files of size 0 from indexing.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

avoid stdio assertion, failing later

remove trailing whitespace

Since mdocml decided to name headers that conflict with system ones (term.h)
move the header inclusion one up.

If you cannot parse .SH NAME, like in the case of the ksh93 man page
where the .SH is followed by a conditional:

.SH NAME
.if \nZ=0 \{\
text text text
.\}

at least don't core-dump.

Make mandb path configurable. makemandb (and related tools) use
the path from the _mandb variable from man.conf now.

Set _mandb in man.conf to same value as was used before.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

branches: 1.14.2;
Use emalloc in one more place, like the rest of the code does.
From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Add -Q flag:
Print only fatal error messages (i.e., when the database is left in
an inconsistent state and needs manual intervention).

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Sync usage with manpage.

Bug fix for PR 46733:
> makemandb always reports the same number for "Total Number of new or
> updated pages enountered" and "Total number of (hard or symbolic)
> links found".

Patch from Abhinav Upadhyay.

Fix typo in a message.

PR 46419 by Abhinav Upadhyay using his updated patch:
Clean up after removing man page aliases.

The new apropos(1) incorrectly displays hyphens in the first line
of the search results for a few man pages (for man(7) based man
pages).

Use patch from Abhinav Upadhyay in PR 46408 to fix this.

branches: 1.7.2;
Fix inverted condition when handling stale entries.
From Abhinav Upadhyay.

Expand workaround for .so usage to do the chdir call just before
starting parsing, not during the tree iteration. This gives it a chance
to work.

Add support for compressed man pages in all the usual formats.

Also handle hyphen replacement if it was used as plain input and no
backslash sequence was used at all in the line.

Be a bit more friendly to man pages using the roff .so command by
changing the current directory to the parent of the man -p entry, e.g.
/usr/share/man for /usr/share/man1.

branches: 1.2.2;
Fix C&P error with $NetBSD$

Import the new apropos/whatis.

This code has been developed by Abhinav Upadhyay as part of Google's Summer
of Code 2011. It uses libmandoc to parse man pages and builds a Full
Text Index in a SQLite database. The combination of indexing the full
manual page, filtering out stop words and ranking individual matches
based on the section gives a much improved user experience.

The old makewhatis and friends are kept under MKMAKEMANDB=no for now.

makemandb/*: fix spelling of database and consistency of SQLite

Don't index outside the mdocs array of function pointers. Analysis and
suggested fixes from Tom Lane. I played it safe and went with (my
variation of) the minimal fix.

port-hppa/56118: sporadic app crashes in HPPA -current

makemandb.c: fail sooner if man page dirs can't be found

There's no point initializing database state if we're then going to
fail to locate any man page sources. Make all the initial state checks
contiguous for simplicity and readability. Also, free the variable
"command" on the error path, and correct the error message.

s/trival/trivial/ in comment.

branches: 1.60.2;
PR misc/54213: Fix performance of whatis(1) when no matches are found

In revision 1.6 of whatis.c the query was modified to return matches for names found
in MLINKS of the man pages as well. However it was slow. The reason probably being that it
required a join. But more importantly the where condition on an FTS virtual table column
is very slow. To avoid the join and the expensive where condition on the virtual table,
add the name_desc column to the mandb_links table as well. This improves the performance
of whatis(1) to the original level at the expense of slight data duplication.

Bump the schema to force database rebuild to take account for the new column addition

remove unneeded header.

adjust to the new mandoc api

Adjust makemandb for the latest mandoc

ok christos@

In the latest mandoc (mdocml) the signature (prototype) of
mparse_alloc() altered - update the call here to compensate.

This fixes the build (of makemandb), but I am not sure that
the changed version is what is desired - someone who knows
something about all of this should validate ... I just copied
the invocation from mandoc's demandoc.c (which seems likely
to be at least a similar kind of usage).

branches: 1.55.8; 1.55.10;
Get rid of unnecessary variable.

We do need to copy the return value from dirname(3) since there it is a static
buffer and can be overwritten in between. I overzealously removed this in one
of my previous commits.

Avoid dereferencing pointer at multiple places, instead use a local variable.

Remove the table name parameter from the check_md5 function.

There is only one table storing the md5 checksums, so we can hardcode the table
name instead of passing it as a function argument.

Avoid copying strings where it is not needed.

Avoid a call to strncmp when comparing only the first character of the string.

Bring the comment in sync with code (after changes brought by the last commit).

Don't parse Nm macro when it occurs anywhere outside the NAME section.

mandoc(3) already generates the text node representing the value for the .Nm macro.
Doing our own parsing for .Nm on top of that leads to large duplication of text
in the database. This gets specially worse for man pages with large NAME sections,
such as queue(3).

branches: 1.47.2;
Use libarchive 3.x interface and not obsolete 2.x versions.

branches: 1.46.2;
Escape hyphen when parsing .Nd

Don't ignore symlinks.
There can be symlinks which are pointing to man pages not installed in
one of the _default locations mentioned in man.conf or MANPATH. For example
there are man pages in /usr/pkg/man which are symlinked to pages in
/usr/pkg/lib/perl5/man. If we ignore symlinks, we would not be able to
index such pages installed outside the default set of directories.

(Also, the symlink test was incorecct, so we never noticed this issue)

Ok christos@, wiz@

We don't need to parse the sections we don't index, so stop early. Saves few
instructions.

With the latest release of mandoc, makemandb(8) started to parse some
sections multiple times. This started to happen because, pmdoc_Sh(), the handler function
responsible for parsing the Sh macros, used to recursively go through all the child
nodes and then the next nodes starting from top level Sh block node.
Now, once it has processed all the child nodes of the top level block node,
it moves to the next node, which is the top level block node of the next section and
in this way one call to pmdoc_Sh() was causing a complete pass through the
man page. Since, mandoc(3) calls pmdoc_Sh() for each .Sh macro in the man
page, it would result in parsing some of the sections multiple times.
This never happened with the previous versions of mandoc, so we never noticed.

I've fixed this by starting the parse sequence of the Sh macro from its body, which gurantees
that we will stop once that section ends.

ok christos@

Use deroff() from mandoc(3) to directly parse the Nd macro rather
than parsing it by hand.

With the latest mandoc(3), the .Nd macro was getting parsed twice. This fixes
that problem and cleans up the code as well.

ok christos@

Fix extraction of section number and machine architecture from man pages.
With the latest API, mdoc_validate()/man_validate() needs to be called before
reading the roff_man.meta field, otherwise it is NULL.

Also, if a man page doesn't specify machine architecture, don't default to '?'
, let it be stored as null in the db. Otherwise, the output of apropos(1) shows
the names of the results as \?/<title>

Sync with API changes.

branches: 1.39.2;
Avoid possible buffer overflow while parsing NAME section of man(7) pages.
Also, simplify copyging of strings, use estrdupn instead of emalloc + memcpy.

Patch from christos@, XXX comment by me

Reuse variable from previous line.

PR/51062: Abhinav Upadhyay: Allow non numeric sections to be indexed and
searched by apropos(1).
Fold long lines.

PR/51040: Abhinav Upadhyay: Fix memory leak

PR/51039: Abhinav Upadhyay: Check for return value of chdir(2)

PR/51034: Abhinav Upadhyay: Close database connection when failed to commit

PR/51034: Abhinav Upadhyay: makemandb(8): Close database connection when
failed to commit

PR/51006: Abhinav Upadhyay: makemandb(8) should parse escape sequences
in the NAME section

Don't crash if we have a missing section.

Adjust to the new mdocml

largely apply patch from PR bin/47392 by Abhinav Upadhyay

change some comments to reflect reality, a variable name to enhance
readability, and adds an assert for safety.

MDOC_MAX is a valid token if the type is text. Adjust.

- handle section numbers that are not single digits
- don't allocate and free needlessly

Explicitly deal with end of lists. PR 49708.

src is too big these days to tolerate superfluous apostrophes. It's
"its", people!

branches: 1.24.2;
Fix a bug that caused an error about a UNIQUE constraint violation.
Patch from Abhinav Upadhyay.

Replace non-breaking space with hyphen, and call hyphen replacement
from one more place.
Improves 'man -k midi' output.

From Abhinav Upadhyay.

branches: 1.22.2;
in update_db(), extract the full list of files to update from the db
before actually updating anything, since changing the db while the query
that extracts the list of files is still in progress results in
the extraction query failing before it finds everything.

Sync with interface change in mdocml 1.12.3.

Skip files of size 0 from indexing.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

avoid stdio assertion, failing later

remove trailing whitespace

Since mdocml decided to name headers that conflict with system ones (term.h)
move the header inclusion one up.

If you cannot parse .SH NAME, like in the case of the ksh93 man page
where the .SH is followed by a conditional:

.SH NAME
.if \nZ=0 \{\
text text text
.\}

at least don't core-dump.

Make mandb path configurable. makemandb (and related tools) use
the path from the _mandb variable from man.conf now.

Set _mandb in man.conf to same value as was used before.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

branches: 1.14.2;
Use emalloc in one more place, like the rest of the code does.
From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Add -Q flag:
Print only fatal error messages (i.e., when the database is left in
an inconsistent state and needs manual intervention).

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Sync usage with manpage.

Bug fix for PR 46733:
> makemandb always reports the same number for "Total Number of new or
> updated pages enountered" and "Total number of (hard or symbolic)
> links found".

Patch from Abhinav Upadhyay.

Fix typo in a message.

PR 46419 by Abhinav Upadhyay using his updated patch:
Clean up after removing man page aliases.

The new apropos(1) incorrectly displays hyphens in the first line
of the search results for a few man pages (for man(7) based man
pages).

Use patch from Abhinav Upadhyay in PR 46408 to fix this.

branches: 1.7.2;
Fix inverted condition when handling stale entries.
From Abhinav Upadhyay.

Expand workaround for .so usage to do the chdir call just before
starting parsing, not during the tree iteration. This gives it a chance
to work.

Add support for compressed man pages in all the usual formats.

Also handle hyphen replacement if it was used as plain input and no
backslash sequence was used at all in the line.

Be a bit more friendly to man pages using the roff .so command by
changing the current directory to the parent of the man -p entry, e.g.
/usr/share/man for /usr/share/man1.

branches: 1.2.2;
Fix C&P error with $NetBSD$

Import the new apropos/whatis.

This code has been developed by Abhinav Upadhyay as part of Google's Summer
of Code 2011. It uses libmandoc to parse man pages and builds a Full
Text Index in a SQLite database. The combination of indexing the full
manual page, filtering out stop words and ranking individual matches
based on the section gives a much improved user experience.

The old makewhatis and friends are kept under MKMAKEMANDB=no for now.

Don't index outside the mdocs array of function pointers. Analysis and
suggested fixes from Tom Lane. I played it safe and went with (my
variation of) the minimal fix.

port-hppa/56118: sporadic app crashes in HPPA -current

makemandb.c: fail sooner if man page dirs can't be found

There's no point initializing database state if we're then going to
fail to locate any man page sources. Make all the initial state checks
contiguous for simplicity and readability. Also, free the variable
"command" on the error path, and correct the error message.

s/trival/trivial/ in comment.

branches: 1.60.2;
PR misc/54213: Fix performance of whatis(1) when no matches are found

In revision 1.6 of whatis.c the query was modified to return matches for names found
in MLINKS of the man pages as well. However it was slow. The reason probably being that it
required a join. But more importantly the where condition on an FTS virtual table column
is very slow. To avoid the join and the expensive where condition on the virtual table,
add the name_desc column to the mandb_links table as well. This improves the performance
of whatis(1) to the original level at the expense of slight data duplication.

Bump the schema to force database rebuild to take account for the new column addition

remove unneeded header.

adjust to the new mandoc api

Adjust makemandb for the latest mandoc

ok christos@

In the latest mandoc (mdocml) the signature (prototype) of
mparse_alloc() altered - update the call here to compensate.

This fixes the build (of makemandb), but I am not sure that
the changed version is what is desired - someone who knows
something about all of this should validate ... I just copied
the invocation from mandoc's demandoc.c (which seems likely
to be at least a similar kind of usage).

branches: 1.55.8; 1.55.10;
Get rid of unnecessary variable.

We do need to copy the return value from dirname(3) since there it is a static
buffer and can be overwritten in between. I overzealously removed this in one
of my previous commits.

Avoid dereferencing pointer at multiple places, instead use a local variable.

Remove the table name parameter from the check_md5 function.

There is only one table storing the md5 checksums, so we can hardcode the table
name instead of passing it as a function argument.

Avoid copying strings where it is not needed.

Avoid a call to strncmp when comparing only the first character of the string.

Bring the comment in sync with code (after changes brought by the last commit).

Don't parse Nm macro when it occurs anywhere outside the NAME section.

mandoc(3) already generates the text node representing the value for the .Nm macro.
Doing our own parsing for .Nm on top of that leads to large duplication of text
in the database. This gets specially worse for man pages with large NAME sections,
such as queue(3).

branches: 1.47.2;
Use libarchive 3.x interface and not obsolete 2.x versions.

branches: 1.46.2;
Escape hyphen when parsing .Nd

Don't ignore symlinks.
There can be symlinks which are pointing to man pages not installed in
one of the _default locations mentioned in man.conf or MANPATH. For example
there are man pages in /usr/pkg/man which are symlinked to pages in
/usr/pkg/lib/perl5/man. If we ignore symlinks, we would not be able to
index such pages installed outside the default set of directories.

(Also, the symlink test was incorecct, so we never noticed this issue)

Ok christos@, wiz@

We don't need to parse the sections we don't index, so stop early. Saves few
instructions.

With the latest release of mandoc, makemandb(8) started to parse some
sections multiple times. This started to happen because, pmdoc_Sh(), the handler function
responsible for parsing the Sh macros, used to recursively go through all the child
nodes and then the next nodes starting from top level Sh block node.
Now, once it has processed all the child nodes of the top level block node,
it moves to the next node, which is the top level block node of the next section and
in this way one call to pmdoc_Sh() was causing a complete pass through the
man page. Since, mandoc(3) calls pmdoc_Sh() for each .Sh macro in the man
page, it would result in parsing some of the sections multiple times.
This never happened with the previous versions of mandoc, so we never noticed.

I've fixed this by starting the parse sequence of the Sh macro from its body, which gurantees
that we will stop once that section ends.

ok christos@

Use deroff() from mandoc(3) to directly parse the Nd macro rather
than parsing it by hand.

With the latest mandoc(3), the .Nd macro was getting parsed twice. This fixes
that problem and cleans up the code as well.

ok christos@

Fix extraction of section number and machine architecture from man pages.
With the latest API, mdoc_validate()/man_validate() needs to be called before
reading the roff_man.meta field, otherwise it is NULL.

Also, if a man page doesn't specify machine architecture, don't default to '?'
, let it be stored as null in the db. Otherwise, the output of apropos(1) shows
the names of the results as \?/<title>

Sync with API changes.

branches: 1.39.2;
Avoid possible buffer overflow while parsing NAME section of man(7) pages.
Also, simplify copyging of strings, use estrdupn instead of emalloc + memcpy.

Patch from christos@, XXX comment by me

Reuse variable from previous line.

PR/51062: Abhinav Upadhyay: Allow non numeric sections to be indexed and
searched by apropos(1).
Fold long lines.

PR/51040: Abhinav Upadhyay: Fix memory leak

PR/51039: Abhinav Upadhyay: Check for return value of chdir(2)

PR/51034: Abhinav Upadhyay: Close database connection when failed to commit

PR/51034: Abhinav Upadhyay: makemandb(8): Close database connection when
failed to commit

PR/51006: Abhinav Upadhyay: makemandb(8) should parse escape sequences
in the NAME section

Don't crash if we have a missing section.

Adjust to the new mdocml

largely apply patch from PR bin/47392 by Abhinav Upadhyay

change some comments to reflect reality, a variable name to enhance
readability, and adds an assert for safety.

MDOC_MAX is a valid token if the type is text. Adjust.

- handle section numbers that are not single digits
- don't allocate and free needlessly

Explicitly deal with end of lists. PR 49708.

src is too big these days to tolerate superfluous apostrophes. It's
"its", people!

branches: 1.24.2;
Fix a bug that caused an error about a UNIQUE constraint violation.
Patch from Abhinav Upadhyay.

Replace non-breaking space with hyphen, and call hyphen replacement
from one more place.
Improves 'man -k midi' output.

From Abhinav Upadhyay.

branches: 1.22.2;
in update_db(), extract the full list of files to update from the db
before actually updating anything, since changing the db while the query
that extracts the list of files is still in progress results in
the extraction query failing before it finds everything.

Sync with interface change in mdocml 1.12.3.

Skip files of size 0 from indexing.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

avoid stdio assertion, failing later

remove trailing whitespace

Since mdocml decided to name headers that conflict with system ones (term.h)
move the header inclusion one up.

If you cannot parse .SH NAME, like in the case of the ksh93 man page
where the .SH is followed by a conditional:

.SH NAME
.if \nZ=0 \{\
text text text
.\}

at least don't core-dump.

Make mandb path configurable. makemandb (and related tools) use
the path from the _mandb variable from man.conf now.

Set _mandb in man.conf to same value as was used before.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

branches: 1.14.2;
Use emalloc in one more place, like the rest of the code does.
From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Add -Q flag:
Print only fatal error messages (i.e., when the database is left in
an inconsistent state and needs manual intervention).

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Sync usage with manpage.

Bug fix for PR 46733:
> makemandb always reports the same number for "Total Number of new or
> updated pages enountered" and "Total number of (hard or symbolic)
> links found".

Patch from Abhinav Upadhyay.

Fix typo in a message.

PR 46419 by Abhinav Upadhyay using his updated patch:
Clean up after removing man page aliases.

The new apropos(1) incorrectly displays hyphens in the first line
of the search results for a few man pages (for man(7) based man
pages).

Use patch from Abhinav Upadhyay in PR 46408 to fix this.

branches: 1.7.2;
Fix inverted condition when handling stale entries.
From Abhinav Upadhyay.

Expand workaround for .so usage to do the chdir call just before
starting parsing, not during the tree iteration. This gives it a chance
to work.

Add support for compressed man pages in all the usual formats.

Also handle hyphen replacement if it was used as plain input and no
backslash sequence was used at all in the line.

Be a bit more friendly to man pages using the roff .so command by
changing the current directory to the parent of the man -p entry, e.g.
/usr/share/man for /usr/share/man1.

branches: 1.2.2;
Fix C&P error with $NetBSD$

Import the new apropos/whatis.

This code has been developed by Abhinav Upadhyay as part of Google's Summer
of Code 2011. It uses libmandoc to parse man pages and builds a Full
Text Index in a SQLite database. The combination of indexing the full
manual page, filtering out stop words and ranking individual matches
based on the section gives a much improved user experience.

The old makewhatis and friends are kept under MKMAKEMANDB=no for now.

makemandb.c: fail sooner if man page dirs can't be found

There's no point initializing database state if we're then going to
fail to locate any man page sources. Make all the initial state checks
contiguous for simplicity and readability. Also, free the variable
"command" on the error path, and correct the error message.

s/trival/trivial/ in comment.

PR misc/54213: Fix performance of whatis(1) when no matches are found

In revision 1.6 of whatis.c the query was modified to return matches for names found
in MLINKS of the man pages as well. However it was slow. The reason probably being that it
required a join. But more importantly the where condition on an FTS virtual table column
is very slow. To avoid the join and the expensive where condition on the virtual table,
add the name_desc column to the mandb_links table as well. This improves the performance
of whatis(1) to the original level at the expense of slight data duplication.

Bump the schema to force database rebuild to take account for the new column addition

remove unneeded header.

adjust to the new mandoc api

Adjust makemandb for the latest mandoc

ok christos@

In the latest mandoc (mdocml) the signature (prototype) of
mparse_alloc() altered - update the call here to compensate.

This fixes the build (of makemandb), but I am not sure that
the changed version is what is desired - someone who knows
something about all of this should validate ... I just copied
the invocation from mandoc's demandoc.c (which seems likely
to be at least a similar kind of usage).

branches: 1.55.8; 1.55.10;
Get rid of unnecessary variable.

We do need to copy the return value from dirname(3) since there it is a static
buffer and can be overwritten in between. I overzealously removed this in one
of my previous commits.

Avoid dereferencing pointer at multiple places, instead use a local variable.

Remove the table name parameter from the check_md5 function.

There is only one table storing the md5 checksums, so we can hardcode the table
name instead of passing it as a function argument.

Avoid copying strings where it is not needed.

Avoid a call to strncmp when comparing only the first character of the string.

Bring the comment in sync with code (after changes brought by the last commit).

Don't parse Nm macro when it occurs anywhere outside the NAME section.

mandoc(3) already generates the text node representing the value for the .Nm macro.
Doing our own parsing for .Nm on top of that leads to large duplication of text
in the database. This gets specially worse for man pages with large NAME sections,
such as queue(3).

branches: 1.47.2;
Use libarchive 3.x interface and not obsolete 2.x versions.

branches: 1.46.2;
Escape hyphen when parsing .Nd

Don't ignore symlinks.
There can be symlinks which are pointing to man pages not installed in
one of the _default locations mentioned in man.conf or MANPATH. For example
there are man pages in /usr/pkg/man which are symlinked to pages in
/usr/pkg/lib/perl5/man. If we ignore symlinks, we would not be able to
index such pages installed outside the default set of directories.

(Also, the symlink test was incorecct, so we never noticed this issue)

Ok christos@, wiz@

We don't need to parse the sections we don't index, so stop early. Saves few
instructions.

With the latest release of mandoc, makemandb(8) started to parse some
sections multiple times. This started to happen because, pmdoc_Sh(), the handler function
responsible for parsing the Sh macros, used to recursively go through all the child
nodes and then the next nodes starting from top level Sh block node.
Now, once it has processed all the child nodes of the top level block node,
it moves to the next node, which is the top level block node of the next section and
in this way one call to pmdoc_Sh() was causing a complete pass through the
man page. Since, mandoc(3) calls pmdoc_Sh() for each .Sh macro in the man
page, it would result in parsing some of the sections multiple times.
This never happened with the previous versions of mandoc, so we never noticed.

I've fixed this by starting the parse sequence of the Sh macro from its body, which gurantees
that we will stop once that section ends.

ok christos@

Use deroff() from mandoc(3) to directly parse the Nd macro rather
than parsing it by hand.

With the latest mandoc(3), the .Nd macro was getting parsed twice. This fixes
that problem and cleans up the code as well.

ok christos@

Fix extraction of section number and machine architecture from man pages.
With the latest API, mdoc_validate()/man_validate() needs to be called before
reading the roff_man.meta field, otherwise it is NULL.

Also, if a man page doesn't specify machine architecture, don't default to '?'
, let it be stored as null in the db. Otherwise, the output of apropos(1) shows
the names of the results as \?/<title>

Sync with API changes.

branches: 1.39.2;
Avoid possible buffer overflow while parsing NAME section of man(7) pages.
Also, simplify copyging of strings, use estrdupn instead of emalloc + memcpy.

Patch from christos@, XXX comment by me

Reuse variable from previous line.

PR/51062: Abhinav Upadhyay: Allow non numeric sections to be indexed and
searched by apropos(1).
Fold long lines.

PR/51040: Abhinav Upadhyay: Fix memory leak

PR/51039: Abhinav Upadhyay: Check for return value of chdir(2)

PR/51034: Abhinav Upadhyay: Close database connection when failed to commit

PR/51034: Abhinav Upadhyay: makemandb(8): Close database connection when
failed to commit

PR/51006: Abhinav Upadhyay: makemandb(8) should parse escape sequences
in the NAME section

Don't crash if we have a missing section.

Adjust to the new mdocml

largely apply patch from PR bin/47392 by Abhinav Upadhyay

change some comments to reflect reality, a variable name to enhance
readability, and adds an assert for safety.

MDOC_MAX is a valid token if the type is text. Adjust.

- handle section numbers that are not single digits
- don't allocate and free needlessly

Explicitly deal with end of lists. PR 49708.

src is too big these days to tolerate superfluous apostrophes. It's
"its", people!

branches: 1.24.2;
Fix a bug that caused an error about a UNIQUE constraint violation.
Patch from Abhinav Upadhyay.

Replace non-breaking space with hyphen, and call hyphen replacement
from one more place.
Improves 'man -k midi' output.

From Abhinav Upadhyay.

branches: 1.22.2;
in update_db(), extract the full list of files to update from the db
before actually updating anything, since changing the db while the query
that extracts the list of files is still in progress results in
the extraction query failing before it finds everything.

Sync with interface change in mdocml 1.12.3.

Skip files of size 0 from indexing.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

avoid stdio assertion, failing later

remove trailing whitespace

Since mdocml decided to name headers that conflict with system ones (term.h)
move the header inclusion one up.

If you cannot parse .SH NAME, like in the case of the ksh93 man page
where the .SH is followed by a conditional:

.SH NAME
.if \nZ=0 \{\
text text text
.\}

at least don't core-dump.

Make mandb path configurable. makemandb (and related tools) use
the path from the _mandb variable from man.conf now.

Set _mandb in man.conf to same value as was used before.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

branches: 1.14.2;
Use emalloc in one more place, like the rest of the code does.
From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Add -Q flag:
Print only fatal error messages (i.e., when the database is left in
an inconsistent state and needs manual intervention).

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

Sync usage with manpage.

Bug fix for PR 46733:
> makemandb always reports the same number for "Total Number of new or
> updated pages enountered" and "Total number of (hard or symbolic)
> links found".

Patch from Abhinav Upadhyay.

Fix typo in a message.

PR 46419 by Abhinav Upadhyay using his updated patch:
Clean up after removing man page aliases.

The new apropos(1) incorrectly displays hyphens in the first line
of the search results for a few man pages (for man(7) based man
pages).

Use patch from Abhinav Upadhyay in PR 46408 to fix this.

branches: 1.7.2;
Fix inverted condition when handling stale entries.
From Abhinav Upadhyay.

Expand workaround for .so usage to do the chdir call just before
starting parsing, not during the tree iteration. This gives it a chance
to work.

Add support for compressed man pages in all the usual formats.

Also handle hyphen replacement if it was used as plain input and no
backslash sequence was used at all in the line.

Be a bit more friendly to man pages using the roff .so command by
changing the current directory to the parent of the man -p entry, e.g.
/usr/share/man for /usr/share/man1.

branches: 1.2.2;
Fix C&P error with $NetBSD$

Import the new apropos/whatis.

This code has been developed by Abhinav Upadhyay as part of Google's Summer
of Code 2011. It uses libmandoc to parse man pages and builds a Full
Text Index in a SQLite database. The combination of indexing the full
manual page, filtering out stop words and ranking individual matches
based on the section gives a much improved user experience.

The old makewhatis and friends are kept under MKMAKEMANDB=no for now.

s/trival/trivial/ in comment.

PR misc/54213: Fix performance of whatis(1) when no matches are found

In revision 1.6 of whatis.c the query was modified to return matches for names found
in MLINKS of the man pages as well. However it was slow. The reason probably being that it
required a join. But more importantly the where condition on an FTS virtual table column
is very slow. To avoid the join and the expensive where condition on the virtual table,
add the name_desc column to the mandb_links table as well. This improves the performance
of whatis(1) to the original level at the expense of slight data duplication.

Bump the schema to force database rebuild to take account for the new column addition

remove unneeded header.

adjust to the new mandoc api

Adjust makemandb for the latest mandoc

ok christos@

In the latest mandoc (mdocml) the signature (prototype) of
mparse_alloc() altered - update the call here to compensate.

This fixes the build (of makemandb), but I am not sure that
the changed version is what is desired - someone who knows
something about all of this should validate ... I just copied
the invocation from mandoc's demandoc.c (which seems likely
to be at least a similar kind of usage).

branches: 1.55.8; 1.55.10;
Get rid of unnecessary variable.

We do need to copy the return value from dirname(3) since there it is a static
buffer and can be overwritten in between. I overzealously removed this in one
of my previous commits.

Avoid dereferencing pointer at multiple places, instead use a local variable.

Remove the table name parameter from the check_md5 function.

There is only one table storing the md5 checksums, so we can hardcode the table
name instead of passing it as a function argument.

Avoid copying strings where it is not needed.

Avoid a call to strncmp when comparing only the first character of the string.

Bring the comment in sync with code (after changes brought by the last commit).

Don't parse Nm macro when it occurs anywhere outside the NAME section.

mandoc(3) already generates the text node representing the value for the .Nm macro.
Doing our own parsing for .Nm on top of that leads to large duplication of text
in the database. This gets specially worse for man pages with large NAME sections,
such as queue(3).

branches: 1.47.2;
Use libarchive 3.x interface and not obsolete 2.x versions.

branches: 1.46.2;
Escape hyphen when parsing .Nd

Don't ignore symlinks.
There can be symlinks which are pointing to man pages not installed in
one of the _default locations mentioned in man.conf or MANPATH. For example
there are man pages in /usr/pkg/man which are symlinked to pages in
/usr/pkg/lib/perl5/man. If we ignore symlinks, we would not be able to
index such pages installed outside the default set of directories.

(Also, the symlink test was incorecct, so we never noticed this issue)

Ok christos@, wiz@

We don't need to parse the sections we don't index, so stop early. Saves few
instructions.

With the latest release of mandoc, makemandb(8) started to parse some
sections multiple times. This started to happen because, pmdoc_Sh(), the handler function
responsible for parsing the Sh macros, used to recursively go through all the child
nodes and then the next nodes starting from top level Sh block node.
Now, once it has processed all the child nodes of the top level block node,
it moves to the next node, which is the top level block node of the next section and
in this way one call to pmdoc_Sh() was causing a complete pass through the
man page. Since, mandoc(3) calls pmdoc_Sh() for each .Sh macro in the man
page, it would result in parsing some of the sections multiple times.
This never happened with the previous versions of mandoc, so we never noticed.

I've fixed this by starting the parse sequence of the Sh macro from its body, which gurantees
that we will stop once that section ends.

ok christos@

Use deroff() from mandoc(3) to directly parse the Nd macro rather
than parsing it by hand.

With the latest mandoc(3), the .Nd macro was getting parsed twice. This fixes
that problem and cleans up the code as well.

ok christos@

Fix extraction of section number and machine architecture from man pages.
With the latest API, mdoc_validate()/man_validate() needs to be called before
reading the roff_man.meta field, otherwise it is NULL.

Also, if a man page doesn't specify machine architecture, don't default to '?'
, let it be stored as null in the db. Otherwise, the output of apropos(1) shows
the names of the results as \?/<title>

Sync with API changes.

branches: 1.39.2;
Avoid possible buffer overflow while parsing NAME section of man(7) pages.
Also, simplify copyging of strings, use estrdupn instead of emalloc + memcpy.

Patch from christos@, XXX comment by me

Reuse variable from previous line.

PR/51062: Abhinav Upadhyay: Allow non numeric sections to be indexed and
searched by apropos(1).
Fold long lines.

PR/51040: Abhinav Upadhyay: Fix memory leak

PR/51039: Abhinav Upadhyay: Check for return value of chdir(2)

PR/51034: Abhinav Upadhyay: Close database connection when failed to commit

PR/51034: Abhinav Upadhyay: makemandb(8): Close database connection when
failed to commit

PR/51006: Abhinav Upadhyay: makemandb(8) should parse escape sequences
in the NAME section

Don't crash if we have a missing section.

Adjust to the new mdocml

largely apply patch from PR bin/47392 by Abhinav Upadhyay

change some comments to reflect reality, a variable name to enhance
readability, and adds an assert for safety.

MDOC_MAX is a valid token if the type is text. Adjust.

- handle section numbers that are not single digits
- don't allocate and free needlessly

Explicitly deal with end of lists. PR 49708.

src is too big these days to tolerate superfluous apostrophes. It's
"its", people!

branches: 1.24.2;
Fix a bug that caused an error about a UNIQUE constraint violation.
Patch from Abhinav Upadhyay.

Replace non-breaking space with hyphen, and call hyphen replacement
from one more place.
Improves 'man -k midi' output.

From Abhinav Upadhyay.

branches: 1.22.2;
in update_db(), extract the full list of files to update from the db
before actually updating anything, since changing the db while the query
that extracts the list of files is still in progress results in
the extraction query failing before it finds everything.

Sync with interface change in mdocml 1.12.3.

Skip files of size 0 from indexing.

From Abhinav Upadhyay <er.abhinav.upadhyay@gmail.com>.

avoid stdio assertion, failing later