Dot leaders are one setting in Writer's Tabs dialog: the row of dots that runs from a job title across to its dates. On this one-page resume, the open-source OpenResume parser went from 38 correct fields to 22 after I added dot leaders. The page looked fine. The dots were in the PDF's text, so the parser read the company, the dots and the city as one long company name.
I started from a Classic ATS-friendly resume template for Writer, with US Letter and A4 downloads, changed one thing at a time in its XML, and exported every version with LibreOffice 26.8.1.1 on macOS. I extracted the text from each PDF with pypdf 7.1.1 and pdfminer.six 20260107. Then I ran OpenResume's parser in Node (commit 4f8255a, pdf.js 3.7.107), with its browser-only PDF loading step swapped for a local copy that maps the text items the same way, and compared the parsed fields with values I'd recorded before the test. I didn't run these PDFs through a commercial ATS such as Workday or Greenhouse. I ran the exports headless, with fontconfig pointed at the macOS system fonts so Writer used Georgia, and I didn't test LibreOffice 7.x.
How the template is built
An .ott file is a Writer template. Opening it starts a new untitled document, and File > Templates > Manage Templates > Import keeps it in your template list. Unzip it, and content.xml shows how the layout is made:
<text:h text:style-name="P3" text:outline-level="2">EXPERIENCE</text:h> <text:p text:style-name="P4"><text:span text:style-name="T1">Harlan Quay Capital Partners</text:span><text:tab/>New York, NY</text:p> The section headings are Heading 2 paragraphs with a 0.74 pt bottom border, so the rule under each heading belongs to the paragraph and isn't a drawn line. The city and the dates sit on the same line as the company and the job title, pushed to the right margin by a right-aligned tab stop at 18.542 cm, the width of the text area on US Letter with 0.6 in margins. The bullets come from a list style whose label is •. There are no tables, frames or page headers.
Exported as it ships, pypdf returned the lines in reading order, with the tab as a single space:
Olivia Bennett Providence, RI | (401) 555-0132 | olivia.bennett@example.com | linkedin.com/in/oliviabennett EDUCATION Halden College Providence, RI Bachelor of Arts in Economics, Minor in Statistics Sep 2022 – May 2026 • GPA: 3.8/4.0, Economics Department Honors Program OpenResume matched all 38 fields. Plain and Student also use right-aligned tab stops and list-style bullets, though their section headings are ordinary paragraphs rather than Writer headings. I ran OpenResume on each template's sample resume. OpenResume matched 33 of 33 fields on the Plain resume template, set in Arial and left-aligned, and 35 of 35 on the Student resume template, which puts education first.
pdfminer.six grouped the same file differently. Its layout analysis put the right-hand cities and dates in blocks of their own and moved some headings below the entries they introduce. In the table, frame and header versions below, pdfminer grouped the text in exactly the same order.
A table didn't change anything here
I expected the table to be the problem. I replaced every tab-stop line with a borderless two-column table: company or title in the left cell, city or dates right-aligned in the right one. pypdf returned exactly the same text as the tab-stop version, character for character. So did pdfminer, and OpenResume still read 38 of 38. With these three readers and Writer 26.8, a plain table was fine. I kept the tab stops, but this test gives no reason to rebuild a table-based resume you already have.
A frame did change the order. I moved the contact line into a frame (Insert > Frame) anchored to a paragraph just below the name. In that version, pypdf put the contact line first and the name second:
Providence, RI | (401) 555-0132 | olivia.bennett@example.com | linkedin.com/in/oliviabennett Olivia Bennett pdfminer sorts by position and kept the name first. OpenResume still matched all 38 fields, but the text-order check in my test setup flagged the contact details as out of order. I also tried moving the name and contact line into a page header (Insert > Header and Footer). pypdf and pdfminer returned the same text as before, with the name and contact line still at the top, and OpenResume still matched all 38 fields. Both moves had one more effect. When I saved the documents as plain text from Writer (the "Text (encoded)" filter, UTF-8), the frame's contact line and everything in the header were missing from the .txt.
Dot leaders and raised ordinals
The dot leaders come from Format > Paragraph > Tabs, where the same right tab stop gets . as its fill character. Writer writes the dots as text, and pypdf returned them:
Harlan Quay Capital Partners...........................................................................................................New York, NY OpenResume read 22 of 38 fields. The company field held the name, the dots and the city. Every job title came back empty, and the date field held the title, the dots and the dates.
Writer 26.8 has ChangeOrdinalNumber on by default, so AutoCorrect can raise the st in 1st as you type. To test the formatting, I added "Placed 1st of 40 teams in a case competition" to a bullet in the XML and applied Writer's superscript formatting to st. pypdf and pdfminer both returned Placed 1st of 40 teams. OpenResume returned Placed 1stof 40 teams. The pdf.js text items show why:
{"str":"Placed 1","x":61.3,"y":385.9,"hasEOL":false} {"str":"st","x":98.9,"y":389.9,"hasEOL":false} {"str":" ","x":103.6,"y":389.9,"hasEOL":false} {"str":"of 40 teams in a case competition; led two workshops on writing clear policy memos for first-year","x":106.1,"y":385.9,"hasEOL":false} The space after st is a pdf.js item of its own, at the same height as the superscript. OpenResume drops whitespace-only items unless they mark a line end, so this space disappears. If AutoCorrect raises an ordinal on your resume, undo right after it happens, set the text back to normal on the Position tab of Format > Character, or turn the option off in the AutoCorrect options dialog.
Some things I expected to matter didn't. Headings typed as "Education" and shown in capitals with Format > Character > Font Effects > Case > UPPERCASE came out as EDUCATION in all three readers, the same as headings typed in capitals. Small capitals did too. I typed • by hand in place of the list bullets. Without the hanging indent, a few lines wrapped differently in pypdf, but the words stayed the same. OpenResume still matched 38 of 38 fields. In pdfminer.six, EDUCATION moved back above the entries.
What Export as PDF adds
In LibreOffice 26.8, Tagged PDF is on by default under File > Export As > Export as PDF. With it on, Writer added a structure tree with H2 tags for the four section headings and list tags for the bullets. (The bookmarks named after the Heading 2 paragraphs come from a separate setting, and they were there with tagging off too.) Tagging also changed one detail of the text. Where a bullet wraps, pypdf and pdfminer both returned a space at the end of the first line. With tagging turned off, the line ended at the last word:
tagged: "... Financial Accounting, Data Analysis " untagged: "... Financial Accounting, Data Analysis" OpenResume read 38 of 38 either way, but a reader that joins wrapped lines without adding a space would get Analysisin from the untagged file. I'd leave Tagged PDF on. PDF/A-2b gave the same text as a normal export. Hybrid PDF embeds the whole .odt inside the PDF. My file grew from 98,918 to 124,861 bytes. I'd send the standard PDF and keep the editable Writer file to myself.
Changing the paper size
I changed the page style from US Letter to A4 and exported another PDF. The text didn't change apart from a few line breaks, and OpenResume still read 38 of 38. But in this template, the right tab stop stayed at 18.542 cm after the switch. The right edge of the dates stayed at 568.8 pt, while the A4 text area ends at 552.1 pt, so the dates ran about 6 mm into the right margin while staying on their own lines. If you need A4, use the A4 download, where the tab stop is at 18 cm, or move the tab stop in every entry paragraph.
A check before you send it
I ran this pypdf check on three of the exports. Install pypdf with python -m pip install pypdf, save the script as check_resume.py, then run python check_resume.py resume.pdf "Olivia Bennett" with your own file name and the name at the top of your resume:
import re import sys from pypdf import PdfReader path, name = sys.argv[1], sys.argv[2] text = "\n".join(page.extract_text() for page in PdfReader(path).pages) lines = [line for line in text.splitlines() if line.strip()] if lines[0].strip() != name: print("first line is not the name:", lines[0][:50]) for line in lines: if re.search(r"(\.|_|-){5,}", line): print("leader characters:", line[:50]) ligatures = re.findall("[\ufb00-\ufb06]", text) if ligatures: print("ligature characters:", ligatures) print("done:", path.rsplit("/", 1)[-1]) On the template as shipped it prints only the done line. On the frame version it prints first line is not the name, and on the dot-leader version it prints ten lines like these:
leader characters: Halden College.................................... leader characters: Bachelor of Arts in Economics, Minor in Statistics It won't catch the raised ordinal, which only OpenResume tripped on. For that, look through the resume for 1st, 2nd and 3rd before you export.
I work on ReadableResume. The person in the sample resume is made up.
I used AI to help draft this post.