Stop using MAXBSIZE to eliminate sys/param.h including (which injects a
ton of namespace intrusion). Create local sizes, and refactor some code
along the way.
ok millert

When expanding $Log$ make sure a newline is added if rd_log has none.
This can happen on new RCS files.

Fixes $Log$ keyword expansion.

Make sure rcs_lock_add() and rcs_lock_remove() properly update the
rd_lock member of the matching delta.

Other keyword substitution will not properly work in cases where the
RCS file did not have a lock for the revision yet or if the lock is
being removed.

Delete unnecessary <libgen.h> #includes

ok deraadt@

Implement Mdocdate keyword substitution.
In particular, this stops rcsdiff(1) and CVSweb from showing
wrong Mdocdate lines in diffs.
Problem originally reported by tj@.
OK tobias@ otto@ millert@

Allow rcsnum_free and buf_free to accept NULL and remove various NULL
checks that are now unnecessary, from Michael W Bombardieri. ok millert

Convert xfree to free. From Fritjof Bornebusch. ok deraadt

Replace <sys/param.h> with <limits.h> and other less dirty headers where
possible. Annotate <sys/param.h> lines with their current reasons. Switch
to PATH_MAX, NGROUPS_MAX, HOST_NAME_MAX+1, LOGIN_NAME_MAX, etc. Change
MIN() and MAX() to local definitions of MINIMUM() and MAXIMUM() where
sensible to avoid pulling in the pollution. These are the files confirmed
through binary verification.
ok guenther, millert, doug (helped with the verification protocol)

Zap some stored values never read; From Fritjof Bornebusch; ok nicm@

ctype cleanups; ok okan

Remove unused variable, from Michael W Bombardieri who also got an ok
from tobias@.

spacing, no binary change.

Replace the keyword expansion code with a faster, more readable version
which additionally also supports $Log$ and $Locker. Written by Joerg
Sonnenberger and tweaked by me.

ok ray

Remove a bunch of unused functions, from Michael W Bombardieri.

ok ray

Remove some unneeded includes and dead code, from Michael W Bombardieri.

ok jasper xsa

previous commit introduced a crash: the removed strtol was being used to
skip past a value and it couldn't just be nuked

- remove unused variables

from Michael W. Bombardieri
ok chl@ nicm@

- plug an fd leak

help/ok chl@ nicm@

Merge -k and expand @@ parsing code between OpenCVS and OpenRCS, resulting
in a performance gain in OpenCVS in some "cvs update" scenarios.

ok zinovik

Merge OpenCVS' rcs_deltatext_set into OpenRCS, fixing a memory leak.

ok nicm zinovik

Remove the need for rp_file in parser structure, instead keep only one
FILE pointer in RCSFILE. This fixes some ugliness in closing an
fdopen()ed FILE and its underlying file descriptor.

Notified by Joerg Sonnenberger <joerg at britannica dot bec to de>

discussed with and ok nicm

Replaced RCS parser code with new rcsparse.{c,h}:
- be very strict about things we parse
- print more information about errors if they occur
- do not fatal() directly in parser, give caller a chance to react
- fix an rcs design issue when it comes to login names

tested by many on tech@

ok xsa

Zapped unused function rcs_tag_resolve.

ok xsa

rcs_state_check must check for more invalid characters than just space.

ok nicm

Ensure that we always parse all deltas before expecting "desc" keyword.

ok nicm, zinovik

Plug memory leak in rcs_getrev, from zinovik.

ok me stsp

Fixed typo.

ok ray, zinovik

rcsnum_parse() allocates already an RCSNUM struct, so it's not needed to
copy it into a newly allocated one (which is even a memory leak).

ok zinovik

style

Make BUFs autoextend by default. This was already done in OpenCVS
a while ago.

OK zinovik nicm

Reduce variable/function name and whitespace differences between
cvs/rcs.

OK xsa zinovik

Rename a bunch of variables and functions, removing the cvs_/rcs_
prefixes to reduce differences between cvs and rcs files.

"yes!" nicm

Only use username if author is NULL. Fixes "ci -wusername" when run
as root.

OK nicm

Don't read before the beginning of the log message if it is zero-length
(possible when it comes from stdin).

Tweaked version of a fix from Peter Hendrickson in PR 6423.

ok deraadt

Make cvs and rcs consistent in what characters they'll accept in
symbol/tag names: require the first character to be a letter and
allow any printable character (except the RCS specials $,.:;@)
afterwards. This matches GNU cvs.

ok xsa@ zinovik@ stsp@

tried to fix a the-the in the comment, ended up rewriting it. with/from joris@

Checking if a file is ASCII should only be done when diffing, because
the non-ASCII characters could be printed to the screen. For
checking in files, checking out files, merging files, and removing
revisions of files, we do so regardless of whether the files are
ASCII or binary.

Fixes PR6031.

OK joris and deraadt.

add our temporary RCS file to our temp_files worklist in case
for some reason we get interrupted before calling rename().

we hate leaving behind temporary files.

Removed unneeded conversion from RCSNUM to string in rcs_rev_setlog.

oki joris@

Support for retrieving revisions from branches in branches was missing.

Reported by chl@, ok joris@

cvs(1) uses a config file to know about the $OpenBSD$ expansion keyword,
but unfortunately rcs is not so lucky, and it will have to encoded inside.
A bit gross, but what can we do? PR 5953, ok tobias

try getlogin() to get the real username when comitting files as root.
this gives better information in log messages when the user logged in
with a different uid and used sudo or su to become root. matches the
behaviour of gcvs except that we don't fallback to LOGNAME or USER
since getlogin() should not lie.

ok tobias@

Memory leak in rcs_delta_stats plugged.

ok joris

From src/usr.bin/cvs:

> add changed lines support to log and rlog (the + and - stuff)
> from Pierre Riteau;

Revert last patch and add this magic branch number only during
output, this way we don't have to hassle in our internal brach handling.

OK joris@

Branches are separated with \n\t in GNU rcs/cvs.
This makes regress tests easier to check with OpenCVS and GNU cvs.

OK xsa@

gnu cvs added a new keyword 'commitid', this diff adds just enough code
to parse the commitid so that cvsweb (using rcs) continues to work on
those files

ok niallo@, xsa@

typo in error msg; from Igor Zinovik.

Rename rcs_diffreg() to diffreg().

OK joris@

Sync revisions and time buffers size to be consistent with each others.
Simplifies further size tweaks if needed.
OK ray@.

value of 'author' in an rcs file is effectively a system username.
therefore, we should allow it to contain the same characters as a username.

pointed out by jcs@
ok ray@ xsa@

Since xrealloc dies on failure it is safe to directly assign to the
original pointer. Theo agrees, and so does the rest of the tree
(ssh, etc. all do this already).

Saves a bunch of variables and assignments.

OK niallo@

Don't free unused variable, from tbert.

OK niallo && xsa

As done in OpenCVS, general includes cleanup sweep. OK otto@.

- fix support for checking out binary files.

testing from xsa@ and Igor Sobrado <igor at string1.ciencias.uniovi.es>
ok xsa@

- plug memleak in rcs_parse_delta()

same as opencvs diff, ok xsa@

Add support to rcsdiff for all the other diff flags that
are pertinent to files (not directories) and that do not
conflict with rcsdiff-specific flags. OK xsa@

Add missing checks for EOF when parsing rcs file sections.
OK xsa@ joris@ niallo@

More strlcpy/strlcat -> snprintf/xasprintf cleanup.

OK joris@

Remove unnecessary strlcpy/strlcat.

``Oh man!'' xsa@

Recalculate buffer length after each expansion.

OK niallo@

Second try at adding binary file support to rcs_expand_keywords(),
using BUFs instead of C strings. Also remove initial keyword search,
which is a premature optimization, slower, and apparently too hard for
me to get right.

OK niallo@

unused variable

spacing

Revert; keywords did not expand.

Add binary file support to rcs_expand_keywords(), using BUFs instead
of C strings.

OK niallo@

fix out of bound read in rcs_expand_keywords(); ok niallo

Remove file mode argument from {cvs|rcs}_buf_write_stmp(). We
always set the temporary file to mode 600, which is already done
by mkstemp().

OK joris@ and xsa@ over a month ago.

- plug an old memory leak in patching code.

ok joris@

- clean up rcs_expand_keywords() a little more.

- zap unused variable

- fix rcs_expand_keywords().

reported by Andreas Kahari <andreas.kahari at gmail>, input from joris@

- explicitly set some pointers to NULL since they are tested on error.

- we can just use the 'sizdiff' variable instead of calling strlen(expbuf);
simplifies things a little.

- ensure fin pointer is not invalidated after realloc

- don't forget to free data in rcs_expand_keywords()

- fix off-by-one in rcs_expand_keywords()

- correctly handle binary files; say bye bye to using c strings for deltatexts.

"slap it in" joris@

Don't fill in the rd_locker if we are creating a new RCS file.

OK joris@

Rename RCSFILE.fd to RCSFILE.rf_fd, plus minor spacing nits. No
binary change.

OK xsa@

first round of rcs_write() cleanup; prodded by joris@; input/OK ray@.

check for rcs_diffreg() return value; OK ray@.

finish work wrt TMPDIR; use xasprintf() to simplify code while in there;
"looks fine" ray@.

- fix a case in rcs_getrev() handling of branches where variant was not being
increased and thus no progress could be made in the loop.

bug noticed in rcsdiff(1) by sturm@

ok joris@

Check return values for all strlcpy, and strlcat calls.

OK xsa@ and probably others.

zap util.[ch] and move the content into rcsutil.[ch].
discussed with joris@.

prevent file races by obtaining an fd for the RCS file and
do our operations on that, this is safe and guarantees we
can operate on the file until we close(2) it.

a fix is coming for the remaining races in our diff code.
okay niallo@ and ray@

err() cleanup;

fork our code we shared between openrcs/cvs into the openrcs dir.

this was starting to become inhuman to maintain without
ugly ugly hacks in the shared code, and it will be easier
to make specific changes for openrcs without touching the
soon-to-be-replaced opencvs code.

When expanding $Log$ make sure a newline is added if rd_log has none.
This can happen on new RCS files.

Fixes $Log$ keyword expansion.

Make sure rcs_lock_add() and rcs_lock_remove() properly update the
rd_lock member of the matching delta.

Other keyword substitution will not properly work in cases where the
RCS file did not have a lock for the revision yet or if the lock is
being removed.

Delete unnecessary <libgen.h> #includes

ok deraadt@

Implement Mdocdate keyword substitution.
In particular, this stops rcsdiff(1) and CVSweb from showing
wrong Mdocdate lines in diffs.
Problem originally reported by tj@.
OK tobias@ otto@ millert@

Allow rcsnum_free and buf_free to accept NULL and remove various NULL
checks that are now unnecessary, from Michael W Bombardieri. ok millert

Convert xfree to free. From Fritjof Bornebusch. ok deraadt

Replace <sys/param.h> with <limits.h> and other less dirty headers where
possible. Annotate <sys/param.h> lines with their current reasons. Switch
to PATH_MAX, NGROUPS_MAX, HOST_NAME_MAX+1, LOGIN_NAME_MAX, etc. Change
MIN() and MAX() to local definitions of MINIMUM() and MAXIMUM() where
sensible to avoid pulling in the pollution. These are the files confirmed
through binary verification.
ok guenther, millert, doug (helped with the verification protocol)

Zap some stored values never read; From Fritjof Bornebusch; ok nicm@

ctype cleanups; ok okan

Remove unused variable, from Michael W Bombardieri who also got an ok
from tobias@.

spacing, no binary change.

Replace the keyword expansion code with a faster, more readable version
which additionally also supports $Log$ and $Locker. Written by Joerg
Sonnenberger and tweaked by me.

ok ray

Remove a bunch of unused functions, from Michael W Bombardieri.

ok ray

Remove some unneeded includes and dead code, from Michael W Bombardieri.

ok jasper xsa

previous commit introduced a crash: the removed strtol was being used to
skip past a value and it couldn't just be nuked

- remove unused variables

from Michael W. Bombardieri
ok chl@ nicm@

- plug an fd leak

help/ok chl@ nicm@

Merge -k and expand @@ parsing code between OpenCVS and OpenRCS, resulting
in a performance gain in OpenCVS in some "cvs update" scenarios.

ok zinovik

Merge OpenCVS' rcs_deltatext_set into OpenRCS, fixing a memory leak.

ok nicm zinovik

Remove the need for rp_file in parser structure, instead keep only one
FILE pointer in RCSFILE. This fixes some ugliness in closing an
fdopen()ed FILE and its underlying file descriptor.

Notified by Joerg Sonnenberger <joerg at britannica dot bec to de>

discussed with and ok nicm

Replaced RCS parser code with new rcsparse.{c,h}:
- be very strict about things we parse
- print more information about errors if they occur
- do not fatal() directly in parser, give caller a chance to react
- fix an rcs design issue when it comes to login names

tested by many on tech@

ok xsa

Zapped unused function rcs_tag_resolve.

ok xsa

rcs_state_check must check for more invalid characters than just space.

ok nicm

Ensure that we always parse all deltas before expecting "desc" keyword.

ok nicm, zinovik

Plug memory leak in rcs_getrev, from zinovik.

ok me stsp

Fixed typo.

ok ray, zinovik

rcsnum_parse() allocates already an RCSNUM struct, so it's not needed to
copy it into a newly allocated one (which is even a memory leak).

ok zinovik

style

Make BUFs autoextend by default. This was already done in OpenCVS
a while ago.

OK zinovik nicm

Reduce variable/function name and whitespace differences between
cvs/rcs.

OK xsa zinovik

Rename a bunch of variables and functions, removing the cvs_/rcs_
prefixes to reduce differences between cvs and rcs files.

"yes!" nicm

Only use username if author is NULL. Fixes "ci -wusername" when run
as root.

OK nicm

Don't read before the beginning of the log message if it is zero-length
(possible when it comes from stdin).

Tweaked version of a fix from Peter Hendrickson in PR 6423.

ok deraadt

Make cvs and rcs consistent in what characters they'll accept in
symbol/tag names: require the first character to be a letter and
allow any printable character (except the RCS specials $,.:;@)
afterwards. This matches GNU cvs.

ok xsa@ zinovik@ stsp@

tried to fix a the-the in the comment, ended up rewriting it. with/from joris@

Checking if a file is ASCII should only be done when diffing, because
the non-ASCII characters could be printed to the screen. For
checking in files, checking out files, merging files, and removing
revisions of files, we do so regardless of whether the files are
ASCII or binary.

Fixes PR6031.

OK joris and deraadt.

add our temporary RCS file to our temp_files worklist in case
for some reason we get interrupted before calling rename().

we hate leaving behind temporary files.

Removed unneeded conversion from RCSNUM to string in rcs_rev_setlog.

oki joris@

Support for retrieving revisions from branches in branches was missing.

Reported by chl@, ok joris@

cvs(1) uses a config file to know about the $OpenBSD$ expansion keyword,
but unfortunately rcs is not so lucky, and it will have to encoded inside.
A bit gross, but what can we do? PR 5953, ok tobias

try getlogin() to get the real username when comitting files as root.
this gives better information in log messages when the user logged in
with a different uid and used sudo or su to become root. matches the
behaviour of gcvs except that we don't fallback to LOGNAME or USER
since getlogin() should not lie.

ok tobias@

Memory leak in rcs_delta_stats plugged.

ok joris

From src/usr.bin/cvs:

> add changed lines support to log and rlog (the + and - stuff)
> from Pierre Riteau;

Revert last patch and add this magic branch number only during
output, this way we don't have to hassle in our internal brach handling.

OK joris@

Branches are separated with \n\t in GNU rcs/cvs.
This makes regress tests easier to check with OpenCVS and GNU cvs.

OK xsa@

gnu cvs added a new keyword 'commitid', this diff adds just enough code
to parse the commitid so that cvsweb (using rcs) continues to work on
those files

ok niallo@, xsa@

typo in error msg; from Igor Zinovik.

Rename rcs_diffreg() to diffreg().

OK joris@

Sync revisions and time buffers size to be consistent with each others.
Simplifies further size tweaks if needed.
OK ray@.

value of 'author' in an rcs file is effectively a system username.
therefore, we should allow it to contain the same characters as a username.

pointed out by jcs@
ok ray@ xsa@

Since xrealloc dies on failure it is safe to directly assign to the
original pointer. Theo agrees, and so does the rest of the tree
(ssh, etc. all do this already).

Saves a bunch of variables and assignments.

OK niallo@

Don't free unused variable, from tbert.

OK niallo && xsa

As done in OpenCVS, general includes cleanup sweep. OK otto@.

- fix support for checking out binary files.

testing from xsa@ and Igor Sobrado <igor at string1.ciencias.uniovi.es>
ok xsa@

- plug memleak in rcs_parse_delta()

same as opencvs diff, ok xsa@

Add support to rcsdiff for all the other diff flags that
are pertinent to files (not directories) and that do not
conflict with rcsdiff-specific flags. OK xsa@

Add missing checks for EOF when parsing rcs file sections.
OK xsa@ joris@ niallo@

More strlcpy/strlcat -> snprintf/xasprintf cleanup.

OK joris@

Remove unnecessary strlcpy/strlcat.

``Oh man!'' xsa@

Recalculate buffer length after each expansion.

OK niallo@

Second try at adding binary file support to rcs_expand_keywords(),
using BUFs instead of C strings. Also remove initial keyword search,
which is a premature optimization, slower, and apparently too hard for
me to get right.

OK niallo@

unused variable

spacing

Revert; keywords did not expand.

Add binary file support to rcs_expand_keywords(), using BUFs instead
of C strings.

OK niallo@

fix out of bound read in rcs_expand_keywords(); ok niallo

Remove file mode argument from {cvs|rcs}_buf_write_stmp(). We
always set the temporary file to mode 600, which is already done
by mkstemp().

OK joris@ and xsa@ over a month ago.

- plug an old memory leak in patching code.

ok joris@

- clean up rcs_expand_keywords() a little more.

- zap unused variable

- fix rcs_expand_keywords().

reported by Andreas Kahari <andreas.kahari at gmail>, input from joris@

- explicitly set some pointers to NULL since they are tested on error.

- we can just use the 'sizdiff' variable instead of calling strlen(expbuf);
simplifies things a little.

- ensure fin pointer is not invalidated after realloc

- don't forget to free data in rcs_expand_keywords()

- fix off-by-one in rcs_expand_keywords()

- correctly handle binary files; say bye bye to using c strings for deltatexts.

"slap it in" joris@

Don't fill in the rd_locker if we are creating a new RCS file.

OK joris@

Rename RCSFILE.fd to RCSFILE.rf_fd, plus minor spacing nits. No
binary change.

OK xsa@

first round of rcs_write() cleanup; prodded by joris@; input/OK ray@.

check for rcs_diffreg() return value; OK ray@.

finish work wrt TMPDIR; use xasprintf() to simplify code while in there;
"looks fine" ray@.

- fix a case in rcs_getrev() handling of branches where variant was not being
increased and thus no progress could be made in the loop.

bug noticed in rcsdiff(1) by sturm@

ok joris@

Check return values for all strlcpy, and strlcat calls.

OK xsa@ and probably others.

zap util.[ch] and move the content into rcsutil.[ch].
discussed with joris@.

prevent file races by obtaining an fd for the RCS file and
do our operations on that, this is safe and guarantees we
can operate on the file until we close(2) it.

a fix is coming for the remaining races in our diff code.
okay niallo@ and ray@

err() cleanup;

fork our code we shared between openrcs/cvs into the openrcs dir.

this was starting to become inhuman to maintain without
ugly ugly hacks in the shared code, and it will be easier
to make specific changes for openrcs without touching the
soon-to-be-replaced opencvs code.

Delete unnecessary <libgen.h> #includes

ok deraadt@

Implement Mdocdate keyword substitution.
In particular, this stops rcsdiff(1) and CVSweb from showing
wrong Mdocdate lines in diffs.
Problem originally reported by tj@.
OK tobias@ otto@ millert@

Allow rcsnum_free and buf_free to accept NULL and remove various NULL
checks that are now unnecessary, from Michael W Bombardieri. ok millert

Convert xfree to free. From Fritjof Bornebusch. ok deraadt

Replace <sys/param.h> with <limits.h> and other less dirty headers where
possible. Annotate <sys/param.h> lines with their current reasons. Switch
to PATH_MAX, NGROUPS_MAX, HOST_NAME_MAX+1, LOGIN_NAME_MAX, etc. Change
MIN() and MAX() to local definitions of MINIMUM() and MAXIMUM() where
sensible to avoid pulling in the pollution. These are the files confirmed
through binary verification.
ok guenther, millert, doug (helped with the verification protocol)

Zap some stored values never read; From Fritjof Bornebusch; ok nicm@

ctype cleanups; ok okan

Remove unused variable, from Michael W Bombardieri who also got an ok
from tobias@.

spacing, no binary change.

Replace the keyword expansion code with a faster, more readable version
which additionally also supports $Log$ and $Locker. Written by Joerg
Sonnenberger and tweaked by me.

ok ray

Remove a bunch of unused functions, from Michael W Bombardieri.

ok ray

Remove some unneeded includes and dead code, from Michael W Bombardieri.

ok jasper xsa

previous commit introduced a crash: the removed strtol was being used to
skip past a value and it couldn't just be nuked

- remove unused variables

from Michael W. Bombardieri
ok chl@ nicm@

- plug an fd leak

help/ok chl@ nicm@

Merge -k and expand @@ parsing code between OpenCVS and OpenRCS, resulting
in a performance gain in OpenCVS in some "cvs update" scenarios.

ok zinovik

Merge OpenCVS' rcs_deltatext_set into OpenRCS, fixing a memory leak.

ok nicm zinovik

Remove the need for rp_file in parser structure, instead keep only one
FILE pointer in RCSFILE. This fixes some ugliness in closing an
fdopen()ed FILE and its underlying file descriptor.

Notified by Joerg Sonnenberger <joerg at britannica dot bec to de>

discussed with and ok nicm

Replaced RCS parser code with new rcsparse.{c,h}:
- be very strict about things we parse
- print more information about errors if they occur
- do not fatal() directly in parser, give caller a chance to react
- fix an rcs design issue when it comes to login names

tested by many on tech@

ok xsa

Zapped unused function rcs_tag_resolve.

ok xsa

rcs_state_check must check for more invalid characters than just space.

ok nicm

Ensure that we always parse all deltas before expecting "desc" keyword.

ok nicm, zinovik

Plug memory leak in rcs_getrev, from zinovik.

ok me stsp

Fixed typo.

ok ray, zinovik

rcsnum_parse() allocates already an RCSNUM struct, so it's not needed to
copy it into a newly allocated one (which is even a memory leak).

ok zinovik

style

Make BUFs autoextend by default. This was already done in OpenCVS
a while ago.

OK zinovik nicm

Reduce variable/function name and whitespace differences between
cvs/rcs.

OK xsa zinovik

Rename a bunch of variables and functions, removing the cvs_/rcs_
prefixes to reduce differences between cvs and rcs files.

"yes!" nicm

Only use username if author is NULL. Fixes "ci -wusername" when run
as root.

OK nicm

Don't read before the beginning of the log message if it is zero-length
(possible when it comes from stdin).

Tweaked version of a fix from Peter Hendrickson in PR 6423.

ok deraadt

Make cvs and rcs consistent in what characters they'll accept in
symbol/tag names: require the first character to be a letter and
allow any printable character (except the RCS specials $,.:;@)
afterwards. This matches GNU cvs.

ok xsa@ zinovik@ stsp@

tried to fix a the-the in the comment, ended up rewriting it. with/from joris@

Checking if a file is ASCII should only be done when diffing, because
the non-ASCII characters could be printed to the screen. For
checking in files, checking out files, merging files, and removing
revisions of files, we do so regardless of whether the files are
ASCII or binary.

Fixes PR6031.

OK joris and deraadt.

add our temporary RCS file to our temp_files worklist in case
for some reason we get interrupted before calling rename().

we hate leaving behind temporary files.

Removed unneeded conversion from RCSNUM to string in rcs_rev_setlog.

oki joris@

Support for retrieving revisions from branches in branches was missing.

Reported by chl@, ok joris@

cvs(1) uses a config file to know about the $OpenBSD$ expansion keyword,
but unfortunately rcs is not so lucky, and it will have to encoded inside.
A bit gross, but what can we do? PR 5953, ok tobias

try getlogin() to get the real username when comitting files as root.
this gives better information in log messages when the user logged in
with a different uid and used sudo or su to become root. matches the
behaviour of gcvs except that we don't fallback to LOGNAME or USER
since getlogin() should not lie.

ok tobias@

Memory leak in rcs_delta_stats plugged.

ok joris

From src/usr.bin/cvs:

> add changed lines support to log and rlog (the + and - stuff)
> from Pierre Riteau;

Revert last patch and add this magic branch number only during
output, this way we don't have to hassle in our internal brach handling.

OK joris@

Branches are separated with \n\t in GNU rcs/cvs.
This makes regress tests easier to check with OpenCVS and GNU cvs.

OK xsa@

gnu cvs added a new keyword 'commitid', this diff adds just enough code
to parse the commitid so that cvsweb (using rcs) continues to work on
those files

ok niallo@, xsa@

typo in error msg; from Igor Zinovik.

Rename rcs_diffreg() to diffreg().

OK joris@

Sync revisions and time buffers size to be consistent with each others.
Simplifies further size tweaks if needed.
OK ray@.

value of 'author' in an rcs file is effectively a system username.
therefore, we should allow it to contain the same characters as a username.

pointed out by jcs@
ok ray@ xsa@

Since xrealloc dies on failure it is safe to directly assign to the
original pointer. Theo agrees, and so does the rest of the tree
(ssh, etc. all do this already).

Saves a bunch of variables and assignments.

OK niallo@

Don't free unused variable, from tbert.

OK niallo && xsa

As done in OpenCVS, general includes cleanup sweep. OK otto@.

- fix support for checking out binary files.

testing from xsa@ and Igor Sobrado <igor at string1.ciencias.uniovi.es>
ok xsa@

- plug memleak in rcs_parse_delta()

same as opencvs diff, ok xsa@

Add support to rcsdiff for all the other diff flags that
are pertinent to files (not directories) and that do not
conflict with rcsdiff-specific flags. OK xsa@

Add missing checks for EOF when parsing rcs file sections.
OK xsa@ joris@ niallo@

More strlcpy/strlcat -> snprintf/xasprintf cleanup.

OK joris@

Remove unnecessary strlcpy/strlcat.

``Oh man!'' xsa@

Recalculate buffer length after each expansion.

OK niallo@

Second try at adding binary file support to rcs_expand_keywords(),
using BUFs instead of C strings. Also remove initial keyword search,
which is a premature optimization, slower, and apparently too hard for
me to get right.

OK niallo@

unused variable

spacing

Revert; keywords did not expand.

Add binary file support to rcs_expand_keywords(), using BUFs instead
of C strings.

OK niallo@

fix out of bound read in rcs_expand_keywords(); ok niallo

Remove file mode argument from {cvs|rcs}_buf_write_stmp(). We
always set the temporary file to mode 600, which is already done
by mkstemp().

OK joris@ and xsa@ over a month ago.

- plug an old memory leak in patching code.

ok joris@

- clean up rcs_expand_keywords() a little more.

- zap unused variable

- fix rcs_expand_keywords().

reported by Andreas Kahari <andreas.kahari at gmail>, input from joris@

- explicitly set some pointers to NULL since they are tested on error.

- we can just use the 'sizdiff' variable instead of calling strlen(expbuf);
simplifies things a little.

- ensure fin pointer is not invalidated after realloc

- don't forget to free data in rcs_expand_keywords()

- fix off-by-one in rcs_expand_keywords()

- correctly handle binary files; say bye bye to using c strings for deltatexts.

"slap it in" joris@

Don't fill in the rd_locker if we are creating a new RCS file.

OK joris@

Rename RCSFILE.fd to RCSFILE.rf_fd, plus minor spacing nits. No
binary change.

OK xsa@

first round of rcs_write() cleanup; prodded by joris@; input/OK ray@.

check for rcs_diffreg() return value; OK ray@.

finish work wrt TMPDIR; use xasprintf() to simplify code while in there;
"looks fine" ray@.

- fix a case in rcs_getrev() handling of branches where variant was not being
increased and thus no progress could be made in the loop.

bug noticed in rcsdiff(1) by sturm@

ok joris@

Check return values for all strlcpy, and strlcat calls.

OK xsa@ and probably others.

zap util.[ch] and move the content into rcsutil.[ch].
discussed with joris@.

prevent file races by obtaining an fd for the RCS file and
do our operations on that, this is safe and guarantees we
can operate on the file until we close(2) it.

a fix is coming for the remaining races in our diff code.
okay niallo@ and ray@

err() cleanup;

fork our code we shared between openrcs/cvs into the openrcs dir.

this was starting to become inhuman to maintain without
ugly ugly hacks in the shared code, and it will be easier
to make specific changes for openrcs without touching the
soon-to-be-replaced opencvs code.

Implement Mdocdate keyword substitution.
In particular, this stops rcsdiff(1) and CVSweb from showing
wrong Mdocdate lines in diffs.
Problem originally reported by tj@.
OK tobias@ otto@ millert@

Allow rcsnum_free and buf_free to accept NULL and remove various NULL
checks that are now unnecessary, from Michael W Bombardieri. ok millert

Convert xfree to free. From Fritjof Bornebusch. ok deraadt

Replace <sys/param.h> with <limits.h> and other less dirty headers where
possible. Annotate <sys/param.h> lines with their current reasons. Switch
to PATH_MAX, NGROUPS_MAX, HOST_NAME_MAX+1, LOGIN_NAME_MAX, etc. Change
MIN() and MAX() to local definitions of MINIMUM() and MAXIMUM() where
sensible to avoid pulling in the pollution. These are the files confirmed
through binary verification.
ok guenther, millert, doug (helped with the verification protocol)

Zap some stored values never read; From Fritjof Bornebusch; ok nicm@

ctype cleanups; ok okan

Remove unused variable, from Michael W Bombardieri who also got an ok
from tobias@.

spacing, no binary change.

Replace the keyword expansion code with a faster, more readable version
which additionally also supports $Log$ and $Locker. Written by Joerg
Sonnenberger and tweaked by me.

ok ray

Remove a bunch of unused functions, from Michael W Bombardieri.

ok ray

Remove some unneeded includes and dead code, from Michael W Bombardieri.

ok jasper xsa

previous commit introduced a crash: the removed strtol was being used to
skip past a value and it couldn't just be nuked

- remove unused variables

from Michael W. Bombardieri
ok chl@ nicm@

- plug an fd leak

help/ok chl@ nicm@

Merge -k and expand @@ parsing code between OpenCVS and OpenRCS, resulting
in a performance gain in OpenCVS in some "cvs update" scenarios.

ok zinovik

Merge OpenCVS' rcs_deltatext_set into OpenRCS, fixing a memory leak.

ok nicm zinovik

Remove the need for rp_file in parser structure, instead keep only one
FILE pointer in RCSFILE. This fixes some ugliness in closing an
fdopen()ed FILE and its underlying file descriptor.

Notified by Joerg Sonnenberger <joerg at britannica dot bec to de>

discussed with and ok nicm

Replaced RCS parser code with new rcsparse.{c,h}:
- be very strict about things we parse
- print more information about errors if they occur
- do not fatal() directly in parser, give caller a chance to react
- fix an rcs design issue when it comes to login names

tested by many on tech@

ok xsa

Zapped unused function rcs_tag_resolve.

ok xsa

rcs_state_check must check for more invalid characters than just space.

ok nicm

Ensure that we always parse all deltas before expecting "desc" keyword.

ok nicm, zinovik

Plug memory leak in rcs_getrev, from zinovik.

ok me stsp

Fixed typo.

ok ray, zinovik

rcsnum_parse() allocates already an RCSNUM struct, so it's not needed to
copy it into a newly allocated one (which is even a memory leak).

ok zinovik

style

Make BUFs autoextend by default. This was already done in OpenCVS
a while ago.

OK zinovik nicm

Reduce variable/function name and whitespace differences between
cvs/rcs.

OK xsa zinovik

Rename a bunch of variables and functions, removing the cvs_/rcs_
prefixes to reduce differences between cvs and rcs files.

"yes!" nicm

Only use username if author is NULL. Fixes "ci -wusername" when run
as root.

OK nicm

Don't read before the beginning of the log message if it is zero-length
(possible when it comes from stdin).

Tweaked version of a fix from Peter Hendrickson in PR 6423.

ok deraadt

Make cvs and rcs consistent in what characters they'll accept in
symbol/tag names: require the first character to be a letter and
allow any printable character (except the RCS specials $,.:;@)
afterwards. This matches GNU cvs.

ok xsa@ zinovik@ stsp@

tried to fix a the-the in the comment, ended up rewriting it. with/from joris@

Checking if a file is ASCII should only be done when diffing, because
the non-ASCII characters could be printed to the screen. For
checking in files, checking out files, merging files, and removing
revisions of files, we do so regardless of whether the files are
ASCII or binary.

Fixes PR6031.

OK joris and deraadt.

add our temporary RCS file to our temp_files worklist in case
for some reason we get interrupted before calling rename().

we hate leaving behind temporary files.

Removed unneeded conversion from RCSNUM to string in rcs_rev_setlog.

oki joris@

Support for retrieving revisions from branches in branches was missing.

Reported by chl@, ok joris@

cvs(1) uses a config file to know about the $OpenBSD$ expansion keyword,
but unfortunately rcs is not so lucky, and it will have to encoded inside.
A bit gross, but what can we do? PR 5953, ok tobias

try getlogin() to get the real username when comitting files as root.
this gives better information in log messages when the user logged in
with a different uid and used sudo or su to become root. matches the
behaviour of gcvs except that we don't fallback to LOGNAME or USER
since getlogin() should not lie.

ok tobias@

Memory leak in rcs_delta_stats plugged.

ok joris

From src/usr.bin/cvs:

> add changed lines support to log and rlog (the + and - stuff)
> from Pierre Riteau;

Revert last patch and add this magic branch number only during
output, this way we don't have to hassle in our internal brach handling.

OK joris@

Branches are separated with \n\t in GNU rcs/cvs.
This makes regress tests easier to check with OpenCVS and GNU cvs.

OK xsa@

gnu cvs added a new keyword 'commitid', this diff adds just enough code
to parse the commitid so that cvsweb (using rcs) continues to work on
those files

ok niallo@, xsa@

typo in error msg; from Igor Zinovik.

Rename rcs_diffreg() to diffreg().

OK joris@

Sync revisions and time buffers size to be consistent with each others.
Simplifies further size tweaks if needed.
OK ray@.

value of 'author' in an rcs file is effectively a system username.
therefore, we should allow it to contain the same characters as a username.

pointed out by jcs@
ok ray@ xsa@

Since xrealloc dies on failure it is safe to directly assign to the
original pointer. Theo agrees, and so does the rest of the tree
(ssh, etc. all do this already).

Saves a bunch of variables and assignments.

OK niallo@

Don't free unused variable, from tbert.

OK niallo && xsa

As done in OpenCVS, general includes cleanup sweep. OK otto@.

- fix support for checking out binary files.

testing from xsa@ and Igor Sobrado <igor at string1.ciencias.uniovi.es>
ok xsa@

- plug memleak in rcs_parse_delta()

same as opencvs diff, ok xsa@

Add support to rcsdiff for all the other diff flags that
are pertinent to files (not directories) and that do not
conflict with rcsdiff-specific flags. OK xsa@

Add missing checks for EOF when parsing rcs file sections.
OK xsa@ joris@ niallo@

More strlcpy/strlcat -> snprintf/xasprintf cleanup.

OK joris@

Remove unnecessary strlcpy/strlcat.

``Oh man!'' xsa@

Recalculate buffer length after each expansion.

OK niallo@

Second try at adding binary file support to rcs_expand_keywords(),
using BUFs instead of C strings. Also remove initial keyword search,
which is a premature optimization, slower, and apparently too hard for
me to get right.

OK niallo@

unused variable

spacing

Revert; keywords did not expand.

Add binary file support to rcs_expand_keywords(), using BUFs instead
of C strings.

OK niallo@

fix out of bound read in rcs_expand_keywords(); ok niallo

Remove file mode argument from {cvs|rcs}_buf_write_stmp(). We
always set the temporary file to mode 600, which is already done
by mkstemp().

OK joris@ and xsa@ over a month ago.

- plug an old memory leak in patching code.

ok joris@

- clean up rcs_expand_keywords() a little more.

- zap unused variable

- fix rcs_expand_keywords().

reported by Andreas Kahari <andreas.kahari at gmail>, input from joris@

- explicitly set some pointers to NULL since they are tested on error.

- we can just use the 'sizdiff' variable instead of calling strlen(expbuf);
simplifies things a little.

- ensure fin pointer is not invalidated after realloc

- don't forget to free data in rcs_expand_keywords()

- fix off-by-one in rcs_expand_keywords()

- correctly handle binary files; say bye bye to using c strings for deltatexts.

"slap it in" joris@

Don't fill in the rd_locker if we are creating a new RCS file.

OK joris@

Rename RCSFILE.fd to RCSFILE.rf_fd, plus minor spacing nits. No
binary change.

OK xsa@

first round of rcs_write() cleanup; prodded by joris@; input/OK ray@.

check for rcs_diffreg() return value; OK ray@.

finish work wrt TMPDIR; use xasprintf() to simplify code while in there;
"looks fine" ray@.

- fix a case in rcs_getrev() handling of branches where variant was not being
increased and thus no progress could be made in the loop.

bug noticed in rcsdiff(1) by sturm@

ok joris@

Check return values for all strlcpy, and strlcat calls.

OK xsa@ and probably others.

zap util.[ch] and move the content into rcsutil.[ch].
discussed with joris@.

prevent file races by obtaining an fd for the RCS file and
do our operations on that, this is safe and guarantees we
can operate on the file until we close(2) it.

a fix is coming for the remaining races in our diff code.
okay niallo@ and ray@

err() cleanup;

fork our code we shared between openrcs/cvs into the openrcs dir.

this was starting to become inhuman to maintain without
ugly ugly hacks in the shared code, and it will be easier
to make specific changes for openrcs without touching the
soon-to-be-replaced opencvs code.