diff options
| author | Paul Buetow <paul@buetow.org> | 2022-02-12 16:11:06 +0000 |
|---|---|---|
| committer | Paul Buetow <paul@buetow.org> | 2022-02-12 16:11:06 +0000 |
| commit | c3388eeefcadda34acc6adf88c9804c27a7ac8af (patch) | |
| tree | 893f7a12248bd399beba9bb41f73c34f7caf68fd | |
| parent | 37df119bec4cc3779132d2c532c84dd42dc8278b (diff) | |
idea about git
| -rw-r--r-- | IDEA.md | 5 | ||||
| -rw-r--r-- | lib/generate.source.sh | 2 |
2 files changed, 5 insertions, 2 deletions
@@ -1,5 +1,9 @@ # IDEAs +## Maybe don't store results in Git + +Just generate the output formats (such as html, md...) and use rsync to push the content to the server. No need to use Git here. Only keep the original .gmi files in git. + ## Parallel job processing queue Currently, in order to speed up, Gemtexter forks on certain functions and loops and joins (via `wait`) on the sub-processes. This however can be a problem once a user max process limit is reached. @@ -10,7 +14,6 @@ Use s.t. like `pgrep -c -P$$` to determine how many sub-processes are already ac I could use pandoc for this (convert from Markdown to PDF). This works on Fedora Linux 34: - ``` sudo dnf install pandoc wkhtmltopdf pandoc FOO.md --pdf-engine=wkhtmltopdf --output FOO.pdf diff --git a/lib/generate.source.sh b/lib/generate.source.sh index c3fc766..d988fde 100644 --- a/lib/generate.source.sh +++ b/lib/generate.source.sh @@ -147,7 +147,7 @@ generate::fromgmi () { # Anoter note: The CNAME file is required by GitHub pages as well for custom domains. for format in "$@"; do find "$CONTENT_BASE_DIR/$format" -type f | - $GREP -E -v '(\.git.*|_config.yml|CNAME|.domains|robots.txt)$'| + $GREP -E -v '(\.git.*|_config.yml|CNAME|.domains|robots.txt|static)$'| while read -r src; do generate::fromgmi_cleanup_docs "$src" "$format" done & |