Feature Request: Support reading gzipped files

[nickman]

GZip encoding can be detected on file open. I realize this may be a performance hit, but it would be extremely helpful.

[variar]

I'd like this for all archive files because in our company logs are attached to bug-reports as zip/7z archives and having to unpack them before opening is pain :) However, I don't see how this can be achived using current line indexing.

[nickbnf]

Indeed, this is highly desirable and I have been thinking about how to do that for a while. Indexing is fine as you can stream the file whilst indexing. The problem is the display, which is done by reading a small, arbitrary part of the file, this is difficult to do on a compressed file unless the compressor as put some marking in it. But I'm still trying to find a solution (which is not to decompress the file fully in memory!!)

[jleaders]

Might I suggest this interface: You are a viewing a compressed file, would you like to decompress it? NOTE: This will disable realtime-streaming of this file, and will decompress in a temporary directory [Decompress]

If you push decompress, do nothing fancy other than unzip to tmp & open that in a new tab (Then popup an error if the tempdir write fails)

Then all these concerns about random access and performance is sidestepped in favor of a one-click, sensible interface. Besides, realtime streaming is rarely needed for compressed files, they are usually archives.

Yes this will use disk space, but since it's by user request, it seems reasonable

[svogt]

I would also vote for decompressing on the fly to a (configurable) temporary directory.

On Linux lograte often times produces gzipped log files (no archive, just gzipped) - I think that's the reason zless / zgrep exist ;)

[variar]

Integrated to latest pre-release klogg build (20.1.0.562, see CI artifacts) a simple version of @jleaders suggestion. zip/7zip/tar.gz/tar.bz2/tar.xz are supported as well as gz/bz2/xz compressed files. Compression is determined by extension. Will use something like libmagic to actually look into file headers in next builds.

Would appreciate any feedback.

[jleaders]

Thanks for your great work! Any hope for it merging down into glogg?

[variar]

There has been no activity in glogg repository since April 2018 :( That is one of the reasons my fork exists.

If @nickbnf is still interested in developing glogg, I will be happy to contribute. If not then finding new maintainer for this great peace of software might be an option.