IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
These are the cases where even the release was not needed (so, the
disttag is not needed, too), or one case where the filename is
contructed (and it doesn't include the disttag).
Now grep -Ee 'headerNVR[^D]' will show the remaining non-trivial
cases, where adapting to disttags may be needed.
- Implemented DistTag support when comparing package versions (with
help by Vladimir D. Seleznev). (It's necessary for the correctness
of checking dependencies after we introduced parseEVRD().)
(Affects rpmbuild, too, when optimizing deps.)
- Dirty hacked to make upgrade packages between branches possible (by
Vladimir D. Seleznev; probably, dead code for rpm-build without rpm-install.)
[rpm-4.13.0.1-alt6 alike]
- rpmEVRcmp() (and hence rpmRangesOverlap()) made asymmetric w.r.t.
underspecified release. (Provides: N = V can't anymore satisfy
Requires: N = V-R. Affects rpmbuild, too, when optimizing deps.)
(with help of Vladimir D. Seleznev)
- rpmRangesOverlap() optimized (can run ca. 30% faster).
In a constraint (Requires, Conflicts), some components of E:V-R:D may
be unspecified; here are the sensible possibilities:
V
E:V-R
E:V-R:D
Remember that the DistTag represents the ID of a particular build. (V
represents a particular upstream version. E:V-R represents a
particular source package release.)
To satisfy a requirement, only the specified components must be
checked.
So, if the requirement doesn't specify a DistTag, then we don't have
to check the DistTag to satisfy it.
If the requirement does specify a DistTag, then if it is an "equals"
requirement, then only the same DistTag can satisfy it. (I.e., we want
that particular build.)
A "less" (or "greater") requirement of a DistTag basically makes no
sense, because the build IDs are not ordered. So, such a requirement
cannot be satisfied, if it has come to checking the DistTag, i.e. the
EVRs have been equal. (It cannot be satisfied by a package with an
equal EVR and any DistTag value, but can be satisfied by a package
with EVR which is strictly less or greater.)
What does the last part mean for "Requires" and "Conflicts"?
Requires: N>E:V-R:D would have the same effect as Requires: N>E:V-R
Conflicts: N>E:V-R:D would have the same effect as Conflicts: N>E:V-R
i.e., a conflict would not disallow packages with the specified E:V-R,
but a different DistTag. (We can't do any better, unless there is a
"not-equals" type of requirement in RPM.)
Commit message author: Ivan Zakharyaschev <imz@altlinux.org>
Commit is based on 6cb615d6112a2ca841481d8153ba652d512a2f23 of
git://git.altlinux.org/gears/r/rpm.git
(cherry picked from commit d169679410a0d02a731addb8b526ecbc8a3a56fc)
This completes the intention of de77ec4556
"build/interdep.c: fix interpackage dependencies automatically":
> ...This change also downgrades "dependency needs Epoch" error to warning
> because it automatically adds strict dependencies.
However, the dependency pkg1->pkg2 was not fixed (if it needed an epoch
or disttag) if pkg1 was already considered to have a strict transitive
dependency on pkg2 (perhaps, via another pkg3: if pkg1->pkg3 and
pkg3->pkg2 deps have already been fixed and added as strict ones).
Then, if pkg3 was a part of a cycle, the pkg1->pkg2 dep wasn't
optimized out. So, it was left untouched, not in the nice format of a
strict dep (i.e., missing an epoch or a disttag).
This was not dangerous, because the strictness was guaranteed by the
transitive dependency, but it was not nice, and was inconvenient for
tools which happen to compare the dep versions as whole strings (say,
for rpm and apt optimizations, or for disttag-unaware old rpm or apt).
- rpmEVRcmp() (and hence rpmRangesOverlap()) made asymmetric w.r.t.
underspecified release. (Provides: N = V can't anymore satisfy
Requires: N = V-R. Look out for unmets!)
(with help of Vladimir D. Seleznev)
- rpmRangesOverlap() optimized (can run ca. 30% faster).
(Analoguous to the change in rpm-4.13-alt6
cfa573f99fbabf7610cec1fb0ee1993f9640b090,
made with help of Vladimir D. Seleznev <vseleznv@altlinux.org>.)
Unchanged (wanted) behavior: On the side of Requires (B), a missing
("underspecified") release means that the relative order of A and B
(result of comparison) is determined only by the speicified components
(epoch, version):
such B is equal to any A with an equal epoch and version (and any release);
such B is greater than A if B's epoch-version is greater than A's;
etc.
A similar treatment of an underspecified release on the side of
Provides (A) was unwanted and it has been changed now:
a B with a non-empty release can't be equal to such A (with a
missing release);
namely, a B with a non-empty release is greater than an A with equal
epoch-version and no release.
Example of a satisfied dependency (worked before and works now):
Provides: N = V-R
Requires: N = V
Example of an unsatisfied dependency (previously, this satisfied the Requires):
Provides: N = V
Requires: N = V-R
We don't want this Requires to be satisfied in this case.
Analoguous strdupa optmization in rpm-4.13-alt6 gave around 30%
improvement in speed when using this functions.
Simplfied code (with variables available only in the scope where they
are used and without extra variables whose value is not used anymore)
is simpler to understand and modify.
- Make "new" packages (with disttags) be treated better
by the "old" disttag-unaware rpm in some cases; primarily those with
< and/or > Conflicts on another subpackage. This form of Conflicts
is used to ensure that no subpackages from different releases/builds
get installed together. (A better way to express this is to add a
common subpackage that all other subpackages depend on.) This change
doesn't affect the way the "new" rpm would treat packages with such
deps (ALT#35930):
+ rewrite < and > dependencies so that they have almost the same meaning when
treated by old disttag-unaware tools;
+ %EVR macro (for intersubpackage deps) upgraded to include %disttag
(given the other change, this is only useful for making the rare
Conflicts: subpkg = %EVR more compatible with disttag-unaware tools).
- checkinstall subpackage added.
Not touched:
Requires: x < E:V-R[:D]
Obsoletes: x > E:V-R[:D]
Requires: x > E:V-R:D (with a disttag)
There are actually no such dependencies in Sisyphus now.
With this change, there are not so many failing tests from
rpminstall-tests-1.0-alt4 (or equivalently from the predisttag branch
there) when run in a system with an "old" rpm and the "new" rpm-build
$ rpm -q rpm --lastchange
* Пт окт 05 2018 Gleb F-Malinovskiy <glebfm@> 4.13.0.1-alt4
- Add _allow_deps_with_beginning_dot macro to allow dependencies
beginning with a dot character in spec file (vseleznv@).
$ ./makeme.sh -j4 SHORT=1
...
XFailed tests:
installable_dummyDisttag_with_reqEqualEpoch (external/strict-old-old dep & future pkg)
noninstallable_dummyDisttag_with_conflEqualEpoch (external/strict-old-old dep & future pkg)
installable_virtDummyDisttag_with_reqEqualEpoch (external/strict-old-old dep & future pkg)
noninstallable_virtDummyDisttag_with_conflEqualEpoch (external/strict-old-old dep & future pkg)
installable_dummyDisttag_with_reqEqual (elusive external/strict-old-old dep & future pkg)
noninstallable_dummyDisttag_with_conflEqual (elusive external/strict-old-old dep & future pkg)
installable_virtDummyDisttag_with_reqEqual (elusive external/strict-old-old dep & future pkg)
noninstallable_virtDummyDisttag_with_conflEqual (elusive external/strict-old-old dep & future pkg)
obsoleted_dummyDisttag_with_obsolEqualEpochDisttag (unrealistic obsoleting disttag)
obsoleted_dummyDisttag_with_obsolEqualDisttag (unrealistic obsoleting disttag)
installable_provVR3Disttag_with_reqEq_VR3 (elusive external/strict-old-old dep & future pkg)
Failed tests:
noninstallable_provOnlyV_with_reqVR
noninstallable_provOnlyV_before_reqVR
installable_provOnlyV_with_conflVR
installable_provOnlyV_before_conflVR
Except for the underspecified Provides problem, the insteresting cases
are the cases of a = dependency.
We can think of a Conflicts: subpkg = E:V-R, which wouldn't work,
but many such cases (if there are any) must be mitigated by the change
of the value of %%EVR to include the disttag (in another commit).
So, the "old" and "new" rpm behave much more similar w.r.t. to the
packages built by the new rpm-build. (This is what we wanted.)
%EVR was introduced to be used in spec-files for intersubpackage
(strict) dependencies.
What if we write (when substituting %%EVR) E:V-R:D into dependencies
instead of the previously written E:V-R? Will the interpretation of
these dependencies by rpm change (when installing packages) in a good
or bad way in any cases?
The "dependencies" can be Requires, Conflicts, Obsoletes.
rpm-build already produces packages which include disttags (D), and
Provides in the form of E:V-R:D, and strict intersubpackage
dependencies of such form (Requires; generated automatically).
Intersubpackage Requires in the form "< %EVR" or "> %EVR" don't make
sense. So, the interesting cases this change can affect are:
* Conflicts of the form "< %EVR" or "> %EVR" or "= %EVR" (the last one
must be quite rare);
* Obsoletes of the form "< %EVR" ("= %EVR" makes little sense;
"> %EVR" looks strange because it's about obsoleting packages from future).
The rpm program can be "new" (which knows how to compare E:V-R:Ds with
disttags; since 4.13.0.1-alt5) or "old" (does not yet know it;
4.13.0.1-alt4 or earlier).
(rpm-build already produces "new" packages, i.e., with disttags.)
The "new" rpm program is considered to be the correct reference way to
treat disttags. According to it, the presence of a disttag in a < or >
dependency cannot affect its interpretation (whether it is satisfied
by another package). Adding a disttag to a = dependency would simply
make it more strict (and that's what rpm-build already does under the
hood for intersubpackage deps). So, E:V-R:D deps are OK for the "new"
rpm.
So, what about about the "old" rpm?
Let's pay attention to the cases where the "old" rpm behaves
differently from the "new" rpm.
Such differences can be demonstrated by running
rpminstall-tests-1.0-alt2.M70P.1 in a system with an "old" rpm and a
"new" rpm-build (that can produce packages with disttags).
Here are the interesting failed test cases without disttag in the dependencies
(like before the change made by this commit), i.e. with dependencies
in the form E:V-R written in Conflicts or Obsoletes:
* installable_dummyDisttag_with_conflGreaterEpoch (and similar ones)
$ rpm -qp RPMS/dummyDisttag/noarch/dummy-1-alt1.noarch.rpm --provides
dummy = 1-alt1:zzz+777.66
$ rpm -qp RPMS/conflGreaterEpoch/noarch/conflGreaterEpoch-1-alt1.noarch.rpm --conflicts
dummy > 0:1-alt1
$ rpm -i RPMS/dummyDisttag/noarch/dummy-1-alt1.noarch.rpm RPMS/conflGreaterEpoch/noarch/conflGreaterEpoch-1-alt1.noarch.rpm
error: Failed dependencies:
dummy > 0:1-alt1 conflicts with conflGreaterEpoch-1-alt1.noarch
NOT OK (ALT#35930): this form of Conflicts is used to ensure that no
subpackages from different releases/builds get installed together.
* noninstallable_dummyDisttag_with_conflEqualEpoch (and similar ones)
$ rpm -qp RPMS/conflEqualEpoch/noarch/conflEqualEpoch-1-alt1.noarch.rpm --conflicts
dummy = 0:1-alt1
$ rpm -i RPMS/dummyDisttag/noarch/dummy-1-alt1.noarch.rpm RPMS/conflEqualEpoch/noarch/conflEqualEpoch-1-alt1.noarch.rpm
succeeds
NOT OK: this form of conflicts could be used to declare conflicting subpackages.
Here are the interesting failed test cases with disttag in the
dependencies (like after the change made by this commit), i.e. with
dependencies in the form E:V-R:D written in Conflicts or Obsoletes:
* installable_dummy_with_conflLessEpochDisttag (and similar ones),
$ rpm -qp RPMS/dummy/noarch/dummy-1-alt1.noarch.rpm --provides
dummy = 1-alt1
$ rpm -qp RPMS/conflLessEpochDisttag/noarch/conflLessEpochDisttag-1-alt1.noarch.rpm --conflicts
dummy < 0:1-alt1:zzz+777.66
$ rpm -i RPMS/dummy/noarch/dummy-1-alt1.noarch.rpm RPMS/conflLessEpochDisttag/noarch/conflLessEpochDisttag-1-alt1.noarch.rpm
error: Failed dependencies:
dummy < 0:1-alt1:zzz+777.66 conflicts with conflLessEpochDisttag-1-alt1.noarch
OK: dummy-1-alt1.noarch.rpm is from another build (if built with the
"new" rpm-build, it would include the disttag in Provides). This form
of Conflicts is used to ensure that no subpackages from different
releases/builds get installed together. The difference in behavior
with the "new" rpm is explained by that the "new" rpm cannot
(unfortunately) detect this subtle mismatch.
* nonobsoleted_dummy_with_obsolLessEpochDisttag (and similar ones)
(like the previous case)
$ rpm -qp RPMS/obsolLessEpochDisttag/noarch/obsolLessEpochDisttag-1-alt1.noarch.rpm --obsoletes
dummy < 0:1-alt1:zzz+777.66
$ rpm -i RPMS/dummy/noarch/dummy-1-alt1.noarch.rpm RPMS/obsolLessEpochDisttag/noarch/obsolLessEpochDisttag-1-alt1.noarch.rpm
$ rpm -q dummy
package dummy is not installed
OK
* nonobsoleted_dummyDisttag_with_obsolLessEpochDisttag (and similar ones)
$ rpm -qp RPMS/dummyDisttag/noarch/dummy-1-alt1.noarch.rpm --provides
dummy = 1-alt1:zzz+777.66
$ rpm -qp RPMS/obsolLessEpochDisttag/noarch/obsolLessEpochDisttag-1-alt1.noarch.rpm --obsoletes
dummy < 0:1-alt1:zzz+777.66
$ rpm -i RPMS/dummyDisttag/noarch/dummy-1-alt1.noarch.rpm RPMS/obsolLessEpochDisttag/noarch/obsolLessEpochDisttag-1-alt1.noarch.rpm
$ rpm -q dummy
package dummy is not installed
NOT OK, but very improbable (the real name of the subpackage must be
the same as the obsoleted name of an older package, but the obsoleting
package is a different subpackage; so, it's improbable that one keeps
the old name of a subpackage, but wants to obsolete it with another
subpackage).
So, introducing this change to the %EVR value would make the new
packages be treated better by the "old" rpm; even the treatment in the
cases where the behavior is different from the behavior of the "new"
rpm is good (except for one improbable case).
getopt(3) holds pointers to previous argument vector in case you want to
continue or restart parsing.
In grabArgs function argument vector is on the stack, so you are lucky
if this memory is not reused for something else between subsequent calls
to the function. If you are not lucky, getopt(3) will read new
arbitrary value from the stack and return error.
The patterns used were OK for the output of "file -NF$'\t'": it would
put a space after the separator (the default separator being ":").
But not for the output of "file -b": we need to pad the result with a
space to use the same patterns.
Putting a space at the beggining is convenient, because it allows to
match independently for "* sh" or "* bash", otherwise "*sh" would
consume "bash", too.
Other uses of "file -b" in scripts/ don't suffer from this problem.
- spec: replaced deprecated PreReq tags with Requires tags.
- Added automatic conversion of deprecated PreReq tags to Requires tags.
- Disallowed extra qualifiers with BuildPreReq tag.
- Disallowed unknown qualifiers with Requires and BuildRequires tags.
- Allowed abbreviated qualifiers with Requires and BuildRequires tags.
- Moved ProvidedSymbols() and SuggestBPP() to separate files.
- lib.prov: Added printing of the number of provided symbols
and the bpp value for each library.
- lib.req: Updated the list of standard libraries with guaranteed versioning.
- suggest_bpp: Fixed harmless off-by-one error in bpp estimation.
ProvidedSymbols() used to be defined both in lib.prov.in and lib.req.in,
fix this code duplication by moving ProvidedSymbols() to separate
provided_symbols executable script.
Likewise, SuggestBPP() used to be defined both in lib.prov.in and
lib.req.in, fix this code duplication by moving SuggestBPP() to separate
suggest_bpp executable script.
Convert PreReq tags without qualifiers into Requires(pre,postun)
and PreReq tags with qualifiers into Requires tags with those qualifiers.
This matches compare_deps() behaviour and opens the way
for additional optimizations of dependencies.
Before this change unknown qualifiers were silently ignored if they
happened to be the last or the only qualifier, e.g. Requires(trash) or
Requires(pre,trash) were accepted but Requires(trash,post) were not.
Now all unknown qualifiers are treated as errors.
Before this change abbreviated qualifiers like BuildRequires(pre)
were silently ignored as unknown, now unambiguously abbreviated qualifiers
are handled like non-abbreviated.
edit_attributes is run twice. Once for phase zero in which all strings are
collected. Then then for phase one in which the strings are rewritten. In
phase zero we also try to collect the comp_dir (either from the
DW_AT_comp_dir or the DW_AT_name of the compile unit). We were also
collecting the comp_dir is phase 1, which is unnecessary, and would not
actually work, since we would be using to old string table index for that,
which had already been rewritten.
Caught by the new string table index checks.
Signed-off-by: Mark Wielaard <mark@klomp.org>
debugedit would blindly use an .debug_str index from the .debug_info or
.debug_line sections assuming it would result in a valid string. Which
would crash and burn if the DWARF data was bogus when the string was
used. So check whenever converting an string index into a char pointer
so we can produce a more helpful error message.
https://bugzilla.redhat.com/show_bug.cgi?id=1543912
Signed-off-by: Mark Wielaard <mark@klomp.org>
edit_dwarf2 calculates the (new) offset in the line program by
taking the difference between the old and new idx, which are of type
size_t (unsigned), plus the size_diff of the header given as ssize_t
(signed), and adding that to the current r_offset, which is an Elf64_Addr
(unsigned). On 64bit architectures, where the size of Elf64_Addr and
ssize_t are the same this isn't a problem. But on 32bit architectures,
where the size of ssize_t is smaller than Elf64_Addr the smaller signed
result gets promoted to an unsigned long first causing issues if the
size_diff was negative.
This would have been caught by gcc -Wsign-conversion
warning: conversion to ‘long unsigned int’ from ‘ssize_t’ {aka ‘long int’}
may change the sign of the result
But enabling this by default gives a lot of false positives.
Found and fixed by Richard Biener <rguenther@suse.de>.
To count as a real directory prefix the string matched should either
be equal to the given prefix or start with the prefix plus '/'.
skip_dir_prefix is always used with base_dir or dest_dir which don't
end with a slash themselves.
This really only is an issue if a package would put a directory named
similar to the package source dir (which cargo on fedora does, by adding
a directory named cargo-vendor in the builddir itself).
Signed-off-by: Mark Wielaard <mark@klomp.org>
The fix for rhbz#444310 (commit c1a5eb - Include empty CU current dirs)
was a little greedy. It would also include comp_dirs outside the build
root. Those are unnecessary and we don't have a good way to store them.
Such dirs (e.g. /tmp) would then show up at the root of /usr/src/debug.
Fix this by including only comp_dirs under base_dir. Also only output
all dirs once (during phase zero) and don't output empty dirs (which
was harmless but would produce a warning from cpio).
This still includes all empty dirs from the original rhbz#444310
nodir testcase and it is an alternative fix for rhbz#641022
(commit c707ab).
Both fixes are necessary in case of an unexpected mode for a directory
actually in the build root that we want to include in the source list.
Signed-off-by: Mark Wielaard <mark@klomp.org>
Some packages depend on the build-ids as generated during the build
and cannot handle rpmbuild recomputing them before generating the
package file list. Add -n, --no-recompute-build-id to debugedit and
add -n to find-debuginfo.sh set by defining the %_no_recompute_build_ids
macro for such packages. %_no_recompute_build_ids can not be used together
with %_unique_build_ids.
Signed-off-by: Mark Wielaard <mark@klomp.org>
We would put one too many slashes in between the new dest_dir and file name
part of the replacement of a DW_FORM_string in the .debug_info. If there
was file part then we would overwrite the first character of the name. If
there was no file part at all then this would overwrite the zero terminator
and cause a crash reading the rest of the data.
A crash did happen while building the docker package on fedora s390x.
https://bugzilla.redhat.com/show_bug.cgi?id=1434347
The reason neither issue would normally trigger is because if we do detect
that the dest_dir is larger than the base_dir we refuse to replace anything.
Signed-off-by: Mark Wielaard <mark@klomp.org>
