| Summary: | app-text/html2text: uses doman on compressed file | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Michał Górny <mgorny> |
| Component: | Current packages | Assignee: | Jaroslav Rakhmatoullin <jazzoslav> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | gmt, proxy-maint |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | 667382 | ||
| Bug Blocks: | 619916 | ||
Thanks. I willh have to read the relevant parts of the package manager spec first. These files are already compressed in the upstream tarball.It seems unnecessary to decompress the man pages if doman supports compressed files.
It this going to change in the future?
The doman helper detects if it is compressed;
31 # These will be automatically decompressed by ecompressdir.
32 if has ${suffix} Z gz bz2 ; then
33 realname=${x%.*}
34 suffix=${realname##*.}
35 fi
This spec says nothing against compressed files:
https://dev.gentoo.org/~ulm/pms/head/pms.html
Which PMS are you referring to?
Should the script simply decompress the files before doman?
(In reply to Jaroslav Rakhmatoullin from comment #2) > These files are already compressed in the upstream tarball.It seems > unnecessary to decompress the man pages if doman supports compressed files. > > It this going to change in the future? Yes. This is supported only in Portage, and it is going to be banned because people keep using the 'it works in Portage' argument and breaking portability. > The doman helper detects if it is compressed; > > > 31 # These will be automatically decompressed by ecompressdir. > 32 if has ${suffix} Z gz bz2 ; then > 33 realname=${x%.*} > 34 suffix=${realname##*.} > 35 fi > > This spec says nothing against compressed files: > > https://dev.gentoo.org/~ulm/pms/head/pms.html > > Which PMS are you referring to? It doesn't say anything about files being encrypted with AES256-CBC-SHA256 using password "bazinga", and encoded in base64 twice either. Nevertheless, even if Portage explicitly accounted for that, would you consider it a good idea to use it? PMS is not some kind of law where you are supposed to look for loopholes. It is a technical specification. If something is not specified, it is undefined. If it is undefined, you can not rely on it for portability. > Should the script simply decompress the files before doman? Yes. (In reply to Michał Górny from comment #3) > It doesn't say anything about files being encrypted with AES256-CBC-SHA256 > using password "bazinga", and encoded in base64 twice either. Nevertheless, > even if Portage explicitly accounted for that, would you consider it a good > idea to use it? still, something to consider for EAPI 8! the only version of this package in the tree decompresses the manpages before calling doman. |
The package uses doman/newman on a compressed file which is invalid per PMS: doman html2text.1.gz html2textrc.5.gz