This fixes a substantial performance issue when scrolling emails with
long/complicated contents, where scrolling down a single line can take
something like hundreds of ms before the screen is updated to reflect
the scroll. It's really bad if the email has lots of columns, e.g. like
if it's an html email that was passed through a filter (w3m, etc) to
render it.
Using pprof, I found that the multiple calls to vterm.ScreenCell.Attrs()
in styleFromCell were really really expensive. This patch replaces them
with a single call.
Here's a before and after with a simple, but very manual test of opening
a large email with contents that went through a w3m filter and
continuously scrolling up and down over and over for ~30 seconds:
*** Before:
----------------------------------------------------------+-------------
flat flat% sum% cum cum% calls calls% + context
----------------------------------------------------------+-------------
28.25s 100% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2fwidgets.PartSwitcher.Draw
0 0% 99.94% 28.25s 82.31% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2fwidgets.PartViewer.Draw
28.25s 100% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2fwidgets.Terminal.Draw
----------------------------------------------------------+-------------
28.25s 100% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2fwidgets.PartViewer.Draw
0 0% 99.94% 28.25s 82.31% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2fwidgets.Terminal.Draw
19.23s 68.07% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2fwidgets.Terminal.styleFromCell
6.04s 21.38% | github.x2ecom..z2fddevault..z2fgo..z2dlibvterm.Screen.GetCellAt
1.38s 4.88% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2flib..z2fui.Context.Printf
0.62s 2.19% | runtime.mapassign
0.43s 1.52% | runtime.mapaccess2
0.20s 0.71% | runtime.newobject
0.19s 0.67% | runtime.callers (inline)
0.07s 0.25% | runtime.makeslice
0.07s 0.25% | runtime.mallocgc
----------------------------------------------------------+-------------
19.23s 100% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2fwidgets.Terminal.Draw
0 0% 99.94% 19.23s 56.03% | git.x2esr.x2eht..z2f..z7esircmpwn..z2faerc..z2fwidgets.Terminal.styleFromCell
19.21s 99.90% | github.x2ecom..z2fddevault..z2fgo..z2dlibvterm.ScreenCell.Attrs
*** After:
----------------------------------------------------------+-------------
flat flat% sum% cum cum% calls calls% + context
----------------------------------------------------------+-------------
0.31s 100% | git.x2esr.x2eht/~sircmpwn/aerc/widgets.Terminal.Draw
0 0% 99.87% 0.31s 1.33% | github.x2ecom/ddevault/go-libvterm.NewPos
0.25s 80.65% | runtime.callers (inline)
0.04s 12.90% | runtime.gomcache (inline)
----------------------------------------------------------+-------------
8.40s 100% | github.x2ecom/ddevault/go-libvterm.Screen.GetCellAt
0 0% 99.87% 8.40s 36.11% | github.x2ecom/ddevault/go-libvterm.Screen.GetCell
7.14s 85.00% | github.x2ecom/ddevault/go-libvterm._cgoCheckPointer
0.54s 6.43% | runtime.callers (inline)
0.35s 4.17% | runtime.exitsyscall
0.11s 1.31% | runtime.deferprocStack
0.07s 0.83% | doentersyscall
0.07s 0.83% | runtime.getg
0.05s 0.6% | runtime.casgstatus
0.03s 0.36% | _init
0.03s 0.36% | runtime.gomcache (inline)
----------------------------------------------------------+-------------
8.46s 100% | git.x2esr.x2eht/~sircmpwn/aerc/widgets.Terminal.Draw
0 0% 99.87% 8.46s 36.37% | github.x2ecom/ddevault/go-libvterm.Screen.GetCellAt
8.40s 99.29% | github.x2ecom/ddevault/go-libvterm.Screen.GetCell
0.06s 0.71% | runtime.callers (inline)
----------------------------------------------------------+-------------
0.31s 100% | git.x2esr.x2eht/~sircmpwn/aerc/widgets.Terminal.styleFromCell
0 0% 99.87% 0.31s 1.33% | github.x2ecom/ddevault/go-libvterm.ScreenCell.Attrs
)
Also update to the tcell v2 PaletteColor api, which should keep the chosen
theme of the user intact.
Note, that if $TRUECOLOR is defined and a truecolor given, aerc will now stop
clipping the value to one of the theme colors.
Generally this is desired behaviour though.
Prior to this commit, the composer was based on a map[string]string.
While this approach was very versatile, it lead to a constant encoding / decoding
of addresses and other headers.
This commit switches to a different model, where the composer is based on a header.
Commands which want to interact with it can simply set some defaults they would
like to have. Users can overwrite them however they like.
In order to get access to the functions generating / getting the msgid go-message
was upgraded.
Allow styles to be layered over a base style. The list of styles to
apply is layered over the base style in order, such that if the layer
does not differ from the base it is not used. The order that these
styles are applied in is, from first to last:
msglist_default
msglist_unread
msglist_read (exclusive with unread, so technically the same level)
msglist_flagged
msglist_deleted
msglist_marked
So, msglist_marked style dominates.
This fixes an issue where the msglist_deleted style was not being applied.
There is a window where a new dir entry isn't yet in the dirlist.dir.
dirlist.ensureScroll however expected to always find a valid index.
Add a check so that we don't try to scroll to a -1 index.
This allows us to hook into the std libs implementation of parsing related stuff.
For this, we need to get rid of the distinction between a mailbox and a host
to just a single "address" field.
However this is already the common case. All but one users immediately
concatenated the mbox/domain to a single address.
So this in effects makes it simpler for most cases and we simply do the
transformation in the special case.
The `:rmdir` command removes the current directory (`-f` is required if
the directory is not empty).
This is not supported on the notmuch backend.
An issue with the maildir backend is that some sync programs (e.g.
offlineimap) may recover the directory after it is deleted. They need
to specifically be configured to accept deletions, or special commands
need to be executed (e.g. `offlineimap --delete-folder`) to properly
delete folders.
A danger of using this on the IMAP backend is that it is possible for a
new message to be added to the directory and for aerc to not show it
immediately (due to a slow connection) - using `:rmdir` at this moment
(with `-f` if the directory already contains messages) would delete the
directory and the new message that just arrived (and all other
contents). This is documented in aerc(1) so that users are aware of
possible risks.
This can happen for example if aerc is compiled without notmuch support but the
notmuch worker is requested.
Pushing a status message isn't good enough, as this gets overridden pretty
quickly if one has multiple accounts configured.
So we show a fullscreen error instead.
Introduce the ability to configure stylesets, allowing customization of
aerc's look (color scheme, font weight, etc). Default styleset is
installed to /path/to/aerc/stylesets/default.
Introduce the ability to configure stylesets, allowing customization of
aerc's look (color scheme, font weight, etc). Default styleset is
installed to /path/to/aerc/stylesets/default.
some people send around huge html without any newline in between.
This did overflow the default 64KB buffer of bufio.Scanner.
If something can't fit in a GB there's no hope left
Also, ignoring errors is bad mkey
Aerc usually used the path []int{1} if it didn't know what the proper path is.
However this only works for multipart messages and breaks if it isn't one.
This patch removes all the hard coding and extracts the necessary helpers to lib.
Added a 'folders-exclude' option that allows removing selected folders
from the directory list sidebar. My motivating example was that removing
a single folder from the list using Golang regexes seemed pretty hard,
so this is a better way to do it. The excluded folders list is included
in the man page.
This changes the scrolling to be done on the draw, when the height is
updated, ensuring that the selected item is kept on screen during
resizing.
Also, this ensures that messages will fill the screen when resizing the
window, for instance, shrinking and then growing drags down more
messages if possible.
This is a transplant of the dirlist scrolling logic.
The grid used static sizes which meant that changing settings didn't
have an effect on elements of the ui, notably the sidebar width. This
patch makes the `Size` parameter of a cell a function which returns the
`int`, allowing for dynamic sizes.
A `Const` function is also included for ease of use for static sizes.
The following functionalities are added to configure aerc ui styles.
- Read stylesets from file with very basic fnmatch wildcard matching
- Add default styleset
- Support different stylesets as part of UiConfig allowing contextual
styles.
- Move widgets/ui elements to use the stylesets.
- Add configuration manual for the styleset
Especially if one tries to interact with all marked messages there could be
the case that not all headers are fetched yet, hence the messageInfo is still nil.
This segfaults a lot of commands which in principle only need the uid to complete.
If we switch to uids, this issue can be alleviated for those commands.
Soves an issue with go1.15 not letting ctty be a parent. See
https://github.com/creack/pty/pull/97 for more details.
Signed-off-by: Guillaume J. Charmes <git+guillaume@charmes.net>
This command uses the Postpone folder from the account config to save
messages to. Messages are saved as though they were sent so have a valid
'to' recipient address and should be able to be read back in for later
editing.
Due to headers being essentially free text, we constantly run into issues
with parts of the body being interpreted as headers.
Remove the ability to overwrite headers to avoid that, while keeping the ability
to specify headers in the template files.
Fixes#383
This adds the commands pin-tab and unpin-tab. Once pinned a tab lives on
the left of the tabstrip and has a configurable marker, defaulting to `
before its name.
This changes the ui to show the spinner while we are sorting. It only
shows one line of the spinner since there are an unknown number of
messages at this time.
Updates to a store can be asynchronous so we shouldn't select it just
because it had an update. Selection of the stores should be driven by
explicit user commands.
This fixes an issue with the updated count logic, where only fetched messages
where counted to the exists string of the rue count.
Note that the count is still broken (we only count read / unread messages we
fetched, but that is the same behaviour as prior to the commit
66b68f35b3)
This reverts commit bd4df53009.
I did not properly untangle the opening / dirlist update of each other.
This interferes with the imap worker, hence the revert
Currently the dirlist ignores the counts provided by the dirInfo.
However some of the workers can actually provide accurate counts much quicker
than if we count the flags.
Eventually we will also want to enable displaying counts for background folders,
where the brute force counting won't work as none of the headers are fetched yet.
This commit models it in an opt-in manner, if the flag isn't set then we still
count the messages manually.
Previously, sending a DirectoryInfo assumed that a directory change
happened. However we don't want that if we only want to update the
unread message count.
The docs of strings.Compare state:
> Compare is included only for symmetry with package bytes. It is usually
> clearer and always faster to use the built-in string comparison operators
> ==, <, >, and so on.
So let's do that.
The go compiler can't help much with untyped int constants.
Even though the only valid constants are 0-3 it will happily accept 4 as input.
Let's let the go compiler worry about correctness here. This also allows people
not very familiar with the code to use it properly via auto completion.
+ Adds parsing of contextual ui sections to aerc config.
+ Add GetUiConfig method for AercConfig that is used to get the
specialized UI config.
+ Add UiConfig method to AccountView to get specialized UI Config.
+ Modifies Aerc codebase to use specialized UIConfig instead.
+ Adds documentation for Contextual UI Configuration
Previously the workers returned a mixture of decoded / encoded parts.
This lead to a whole bunch of issues.
This commit changes the msgviewer and the commands to assume parts to already
be decoded
In order to accommodate for that, the headerlayout needed to be rewritten
to pass the filter criteria back to the msgviewer, instead of just using
the normal headers.
Note that, until we get color configuration, this means that the user *must*
have the %Z verb in the index format else it'll be horribly confusing
as no visual indication is provided
