mod_zip
mod_zip copied to clipboard
evanmiller
Streaming ZIP archiver for nginx 📦
Currently, every time mod_zip reads 8KB, it allocates a new string, copies all existing data, and doesn't free the old string. This results in GBs of memory use for a...
Currently, mod_zip [decodes the path after parsing it](https://github.com/evanmiller/mod_zip/blob/master/ngx_http_zip_parsers.rl#L120), only for Nginx to encode it again when making a sub-request. There are two different but related problems here: - mod_zip decodes...
This causes two problems: - '+' gets incorrectly decoded to ' ' - '%2B' gets decoded to '+', but not encoded back to '%2B', resulting in a slightly different URL...
Hello, I'm trying to configure **NGINX** with this dynamic module inside a **Docker container**, and in [this post](https://www.reddit.com/r/nginx/comments/9jzqa6/trying_to_understand_mod_zip/) I found a way (maybe, but I'm not sure) to implement **mod_zip**...
We use nginx 1.22.1 with version 1.3.0 and see excessive memory usage. We create zip files with local files only based on a manifest without CRC checksums. For a small...
Hello everyone! ``` uwsgi_intercept_errors on; upstream geoserver_proxy { server 10.0.14.84:8080; } upstream ziplist { server 10.0.14.84:5555; } # Expires map map $sent_http_content_type $expires { default off; text/html epoch; text/css max;...
An empty zip file is completely valid, but the current mod_zip syntax does not support them. This comes up in cases where e.g. a user clicks "Download this folder" -...
Related to #74. Generating consistent timestamp is better than having some arbitrary and changing timestamp. It's also weird when dealing with resumed download as the timestamps will be all over...
CRC32 values are now correctly stored for upstream requests.
