<?xml version="1.0" encoding="UTF-8" ?>

<bugzilla version="5.2"
          urlbase="https://bugzilla.altlinux.org/"
          
          maintainer="jenya@basealt.ru"
>

    <bug>
          <bug_id>213</bug_id>
          
          <creation_ts>2001-12-08 22:05:47 +0300</creation_ts>
          <short_desc>extendable catalogue for buildreq of what should be ignored</short_desc>
          <delta_ts>2003-08-25 15:18:25 +0400</delta_ts>
          <reporter_accessible>1</reporter_accessible>
          <cclist_accessible>1</cclist_accessible>
          <classification_id>4</classification_id>
          <classification>Development</classification>
          <product>Sisyphus</product>
          <component>rpm-utils</component>
          <version>unstable</version>
          <rep_platform>all</rep_platform>
          <op_sys>Linux</op_sys>
          <bug_status>CLOSED</bug_status>
          <resolution>FIXED</resolution>
          
          
          <bug_file_loc></bug_file_loc>
          <status_whiteboard></status_whiteboard>
          <keywords></keywords>
          <priority>P5</priority>
          <bug_severity>enhancement</bug_severity>
          <target_milestone>---</target_milestone>
          
          
          <everconfirmed>1</everconfirmed>
          <reporter name="imz">vanyaz</reporter>
          <assigned_to name="placeholder@altlinux.org">placeholder</assigned_to>
          <cc>at</cc>
    
    <cc>glebfm</cc>
    
    <cc>ldv</cc>
    
    <cc>placeholder</cc>
    
    <cc>vt</cc>
          
          

      

      

      

          <comment_sort_order>oldest_to_newest</comment_sort_order>  
          <long_desc isprivate="0" >
    <commentid>202</commentid>
    <comment_count>0</comment_count>
    <who name="imz">vanyaz</who>
    <bug_when>2001-12-08 22:05:47 +0300</bug_when>
    <thetext>Other packages should be able to extend the rules that define which files are essential for detecting dependencies.

For example, emacs-common package could install an additional rule saying that /etc/emacs/site-start.d/* must be ignored. Or python could install a single rule saying that /etc/emacs/site-start.d/python.el must be ignored.

This would make this tool more flexible, and the maintainer of buildreq would not require to add each and other unessential file to the list.

---

---
This applies to rpm-utils-0.2.3-alt1.
</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>2827</commentid>
    <comment_count>1</comment_count>
    <who name="Dmitry V. Levin">ldv</who>
    <bug_when>2001-12-10 18:05:56 +0300</bug_when>
    <thetext>There is a file already used for this purposes:
/etc/buildreqs/files/ignore

Or maybe you mean user-defined file, for example, ~/.buildreqs/files/ignore?

In the last case, should it just override system ignore file?
</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>6192</commentid>
    <comment_count>2</comment_count>
    <who name="Dmitry V. Levin">ldv</who>
    <bug_when>2001-12-10 18:05:56 +0300</bug_when>
    <thetext>There is a file already used for this purposes:
/etc/buildreqs/files/ignore

Or maybe you mean user-defined file, for example, ~/.buildreqs/files/ignore?

In the last case, should it just override system ignore file?
</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>2828</commentid>
    <comment_count>3</comment_count>
    <who name="imz">vanyaz</who>
    <bug_when>2001-12-10 19:20:38 +0300</bug_when>
    <thetext>I mean /etc/buildreqs/files/ignore.d/  ;-)

(Anyway, probably now before Master release it\&apos;s not the right time for talking about this and impelementing it; so the long comment can be postponed.)

So that every package that is responsible for some structures in the filesystem that generate \&amp;quot;wrong\&amp;quot; dependencies installs a file into this directory; this file contains an additional rule (pattern) describing the files to ignore.

For example, emacs-common could install /etc/buildreqs/files/ignore.d/emacs, where it is written to ignore /etc/emacs/site-start.d/*. (Because the packager of emacs-common knows how it uses these scripts, he knows that they are not essential, but the maintainer of buildreq may have no idea about them.)

Or, a more detailed variant is possible when each package installs a list of its own files which must be ignored, which must not generate a dependency on this package. In the example with Emacs, python would install the file /etc/buildreqs/files/ignore.d/python that contains \&amp;quot;/etc/emacs/site-start.d/python.el\&amp;quot;. This would cause buildreq not to take \&amp;quot;pyhton.el\&amp;quot; into account when generating dependency list.


The last (detailed) scheme would require a rather large amount of data to be processed by buildreq, and probably it would be better to implement it with a real (not plain text) database, or even integrate it with RPM so that a packager can write something like this in the spec file (python.spec):

%files
%ignore_for_buildreq %config(noreplace) /etc/emacs/site-start.d/python.el
.....

and then when pyhton package is installed, the corresponding information is installed into the database for buildreq of ignored files.

</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>6193</commentid>
    <comment_count>4</comment_count>
    <who name="imz">vanyaz</who>
    <bug_when>2001-12-10 19:20:38 +0300</bug_when>
    <thetext>I mean /etc/buildreqs/files/ignore.d/  ;-)

(Anyway, probably now before Master release it\&apos;s not the right time for talking about this and impelementing it; so the long comment can be postponed.)

So that every package that is responsible for some structures in the filesystem that generate \&amp;quot;wrong\&amp;quot; dependencies installs a file into this directory; this file contains an additional rule (pattern) describing the files to ignore.

For example, emacs-common could install /etc/buildreqs/files/ignore.d/emacs, where it is written to ignore /etc/emacs/site-start.d/*. (Because the packager of emacs-common knows how it uses these scripts, he knows that they are not essential, but the maintainer of buildreq may have no idea about them.)

Or, a more detailed variant is possible when each package installs a list of its own files which must be ignored, which must not generate a dependency on this package. In the example with Emacs, python would install the file /etc/buildreqs/files/ignore.d/python that contains \&amp;quot;/etc/emacs/site-start.d/python.el\&amp;quot;. This would cause buildreq not to take \&amp;quot;pyhton.el\&amp;quot; into account when generating dependency list.


The last (detailed) scheme would require a rather large amount of data to be processed by buildreq, and probably it would be better to implement it with a real (not plain text) database, or even integrate it with RPM so that a packager can write something like this in the spec file (python.spec):

%files
%ignore_for_buildreq %config(noreplace) /etc/emacs/site-start.d/python.el
.....

and then when pyhton package is installed, the corresponding information is installed into the database for buildreq of ignored files.

</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>2829</commentid>
    <comment_count>5</comment_count>
    <who name="Dmitry V. Levin">ldv</who>
    <bug_when>2002-03-18 22:16:41 +0300</bug_when>
    <thetext>/etc/buildreqs/files/ignore.d
implemented in
rpm-utils-0.3.0-alt1
</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>6194</commentid>
    <comment_count>6</comment_count>
    <who name="Dmitry V. Levin">ldv</who>
    <bug_when>2002-03-18 22:16:41 +0300</bug_when>
    <thetext>/etc/buildreqs/files/ignore.d
implemented in
rpm-utils-0.3.0-alt1
</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>2830</commentid>
    <comment_count>7</comment_count>
    <who name="Dmitry V. Levin">ldv</who>
    <bug_when>2002-09-28 21:32:45 +0400</bug_when>
    <thetext>The ignore.d solution seems to be ok; if not, reopen.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>6195</commentid>
    <comment_count>8</comment_count>
    <who name="Dmitry V. Levin">ldv</who>
    <bug_when>2002-09-28 21:32:45 +0400</bug_when>
    <thetext>The ignore.d solution seems to be ok; if not, reopen.</thetext>
  </long_desc>
      
      

    </bug>

</bugzilla>